USF Libraries
USF Digital Collections

Luminex microsphere immunoassay offers an improved method in testing for antibodies to Eastern Equine Encephalitis virus...


Material Information

Luminex microsphere immunoassay offers an improved method in testing for antibodies to Eastern Equine Encephalitis virus in sentinel chicken sera
Physical Description:
Fitzpatrick, Kelly Ann
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:


Subjects / Keywords:
Dissertations, Academic -- Global Health -- Masters -- USF   ( lcsh )
non-fiction   ( marcgt )


ABSTRACT: Eastern Equine Encephalitis virus has a human mortality rate of 30% of those cases diagnosed, while 30% of those surviving infection remain with neurological sequelae for life (, 2007). The use of sentinel chickens for surveillance of arboviruses that are known to use birds as a reservoir host, such as St. Louis Encephalitis (SLE), West Nile (WN) virus, Eastern Equine Encephalitis (EEE) and Highlands J (HJ) virus, in Florida began with the Sentinel Chicken Arboviral Surveillance Network in 1978 (Day and Stark, 1996). This network enables the activation of an early warning system for citizens, as well as, county epidemiologists and those in mosquito control, allowing for a coordinated effort of disease prevention.Methods currently used at the Florida Department of Health, Tampa Branch Laboratory include screening of submitted sera for antibodies to these arboviruses of epidemiologic importance by way of the hemagglutination inhibition test (HAI), and confirmation by the IgM antibody capture enzyme-linked immunosorbent assay (MAC-ELISA) and Plaque Reduction Neutralization test if the MAC-ELISA proves to be negative. While these tests combined are providing the results needed, the time to result can be a week or greater depending on the initial screening result in the HAI tests. The Microsphere Assay Technology provides an accurate, more rapid (a day or two instead of a week or more) detection method including both a screening and confirmation protocol specifically designed to test for antibody to EEE in sentinel chicken sera.Two sera out of the thousands tested that were tested by HAI shown to be negative in standard testing, resulted as positive by the MIA method and therefore indicated a missed positive. The sensitivity and specificity, positive and negative predictive values of this new protocol as compared with MAC-ELISA as a reference standard indicated that both tests were remarkably similar; Providing sensitivity near 80%, specificity and PPV at 99%, and negative predictive values at 90% for MAC-ELISA and 94% for the MIA. Finally it was determined that Highlands J virus will not have any impact on the testing protocol and results of this test.
Thesis (M.S.P.H.)--University of South Florida, 2008.
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Kelly Ann Fitzpatrick.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 103, 4 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002000377
oclc - 318896526
usfldc doi - E14-SFE0002480
usfldc handle - e14.2480
System ID:

This item is only available as the following downloads:

Full Text


Luminex Microsphere Immunoassay Offers an Improved Method in Testing for Antibodies to Eastern Equine Encephalitis Virus in Sentinel Chicken Sera by Kelly Ann Fitzpatrick A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Public Health Department of Global Health College of Public Health University of South Florida Major Professor: Lillian M. Stark, Ph.D. Jacqueline Cattani, Ph.D. Azliyati Azizan, Ph.D. Date of Approval: July 18, 2008 Keywords: flavivirus, alphavirus, arboviruses, surv eillance, MAC-ELISA Copyright 2008, Kelly Ann Fitzpatrick


Dedication To my family who provided the encouragement, argume nt, love and attitude to make my goals a reality, and to my friends who leant their shoulders in support of my mentality.


Acknowledgements Completing a Masters Degree at the University of So uth Florida, College of Public Health has been a phenomenal experience. I have lea rned a great deal during this time about teamwork, application of knowledge and even m yself and my limits. I am grateful to have been provided the great opportunity in purs uing this research by the Department of Health, Florida Bureau of Laboratories. In parti cular I wish to thank the extraordinary ladies on my thesis committee, Jacqueline Cattani P h.D., Azliyati Azizan Ph.D., and especially my mentor Lillian Stark Ph.D., MPH for p roviding me support and inspiration for the questions and methods behind this new testi ng protocol. In addition I wish to thank Maribel Casteneda and Rita Judge for initiati ng me onto the path of understanding the Arbovirus testing methods currently in place at the Department of Health, including the Luminex assay originally designed by Logan Hall er MSPH. Thank you to Ann Mutilinsky for her generous contribution of time to the MAC-ELISA results, without which there would have been no standard of referenc e for my cutoff values. Of course also I wish to mention and thank Deno Kazanis Ph.D. and Christy Ottendorfer Ph.D. for their much needed experience and willingness to gui de me through my doubts, design and deliberations. To Angela Butler and Jazmine Mat eus, a thank you for your time and patience helping me with the statistical analysis a nd SAS; finally a thank you to all of the people at the College of Public Health and the Bure au of Laboratories for making this an experience of a lifetime.


i Table of Contents List of Tables iii List of Figures v List of Symbols and Abbreviations vii Abstract viii Introduction 1 Arthropod-Borne viruses 1 Alphaviruses 8 Epidemiology and Pathogenesis 8 Classification and Antigenic Types 9 Immunological Response 10 Eastern Equine Encephalitis 12 History 12 Epidemiology 14 Host Vector and Transmission 16 Clinical Features 18 Humans 17 Horses and Other Ungulates 19 Other Mammals 22 Fowl 24 Treatment and Prevention 25 Arbovirus Surveillance 26 Florida Sentinel Chicken Program 28 Serological Antibody Detection Method 29 Hemagglutination Inhibition Test 30 IgM Antibody Capture Enzyme Linked Immunosorbent Assay 32 Plaque Reduction Neutralization Test 35 Microsphere Immunoassays 36 Objectives 39 Materials and Methods 41 Processing and submission of Sentinel Chicken Sera Samples 41


ii Sample Size and Specimen Selection 45 Serum Analysis 45 Microsphere-based Immunoassay 51 Addition of Antigen to Bead Sets 51 Standardization of Microsphere-Based Immunoassay 53 Microsphere-Based Immunoassay Protocol 57 Discernment of IgM Antibodies to EEE 58 Classification of Luminex MFI Results 61 Results 63 Luminex Microsphere Immunoassay Technology 63 IgG Depletion of Sentinel Chicken Sera 64 Dilution Factors for Testing Protocol 66 Classification of the Microsphere–Based Immunoass ay Result 75 Detection of Antibodies to Eastern Equine Encepha litis 77 HAI Negative Samples Showing Luminex Positivity 79 Assay Sensitivity, Specificity, Positive Predictiv e Value, and Negative Predictive Values 82 Testing Results of the cross-reactive Alphavirus H ighlands J 84 Discussion 85 Summary and Conclusions 101 References 104 Bibliography 108


iii List of Tables Table 1 Arthropod-borne viral families of significa nce 4 Table 2 Total sera samples submitted during the 200 7-year and samples tested in this research both regionalized by locat ion 43 Table 3 Sentinel chicken seroconversion rates by re gion and state for 2002-2007 46 Table 4 Counties by region used for sample selectio n 47 Table 5 The sample size for each region as determin ed using the sample size calculator from Cameron and Baldock (1 998) 48 Table 6 Volume of reagents used to create antigen coupled bead set 15 at 500 beads/l with a total volume of 1000 m l 52 Table 7 Combinations employed to confirm absence o f reactivity in incomplete wells. 56 Table 8 Single Factor ANOVA table for depleted vs. non-depleted Chicken sera 65 Table 9 Single Factor ANOVA table for bead set dilu tions at sera dilutions of 1:80, 1:160, 1:320, 1:400, 1:640 and 1 :1280 67 Table 10 Tukey’s multiple comparison test for the bead set dilutions 1:5, 1:10 and 1:20, calculated from P/N transform ed data with positive and negative antigen. 68 Table 11a Tukey Comparisons and grouping for signi ficance of difference between various Ab dilutions using P/N transformed MFI values 71


iv Table 11b. Tukey comparison groupings for signific ance of difference between various Ab dilutions using P/N transformed MFI values. 72 Table 12 Sera found negative by conventional metho ds, found positive in the MIA. 81 Table 13 Sensitivity, Specificity, Positive and Ne gative predictive values for EEE MIA protocol vs. the HAI and ELISA method s 85 Table 14 Comparison of testing methods to the true value in 2X2 tables. 95


v List of Figures Figure 1 Diagram of the sylvatic and peridomestic c ycles of Eastern Equine Encephalitis 13 Figure 2 Flowchart for sentinel chicken sera testin g for antibody to alphaviruses at the Florida Department of Health Bu reau of Laboratories 44 Figure 3 Confirmation Plate design used for testing of chicken sera for IgM antibodies 60 Figure 4 Column graph of primary and secondary ant ibody dilution with viral antigen positive bead sets 73 Figure 5 Column graph of primary and secondary ant ibody dilution with EEE negative control bead sets. 74 Figure 6 EEE ROC curve showing a visual representa tion of the cutoff value (9.7). 76 Figure 7 Comparison of sensitivity and specificity for the Hemagglutination Inhibition test (HAI), MAC-ELISA, and the Microsphere-based Immuno assay (MIA) for the detect ion of antibodies to Eastern Equine Encephalitis virus (EE E) compared to the true value results 96 Figure 8 Comparison of Positive and Negative predi ctive values (PPV, NPV) for the Hemagglutination Inhibition tes t (HAI), MAC-ELISA, and the Microsphere-based Immuno assay (MIA) for the detection of antibodies to Eastern Equine Encephalitis virus (EEE) compared to the true value results 97 Figure 9 Comparison of sensitivity and specificity for the Hemagglutination Inhibition test (HAI), MAC-ELISA, against the screening and confirmation results of the Microsphere-based Immuno assay (MIA) for the detec tion of antibodies to Eastern Equine Encephalitis virus (E EE 98


vi Figure 10 Comparison of Positive and Negative pred ictive values (PPV, NPV) for the Hemagglutination Inhibition te st (HAI), MAC-ELISA, against the screening and confirmation results of the Microsphere-based Immuno assay (MIA) for t he detection of antibodies to Eastern Equine Encephalitis viru s (EEE). 99 Figure 11 MIA final results incorporating both the screening and confirmation techniques for one final result vs. the IgM ELISA, the current gold standard for IgM detectio n in sentinel chicken sera. 100


vii List of Symbols and Abbreviations Symbol and Abbreviations Description % Percent C Degrees Centigrade Ab Antibody Ag Antigen BABS Bovine Albumin-Borate Saline CDC Centers for Disease Control and Prevention CNS Central Nervous System CSF Cerebrospinal Fluid EEEV Eastern Equine Encephalomyelitis Virus EIA ELISA Immunoassay FBE Florida Bureau of Epidemiology FDOH Florida Department of Health g gravity HAI Hemagglutination Inhibition Assay HJV Highlands J Virus IgG Immunoglobulin G IgM Immunoglobulin M JE Japanese Encephalitis Virus MFI Mean Fluorescent Intensity MIA Microsphere-based Immunoassay MAC-ELISA IgM Antibody Capture Enzyme-Linked mRNA Messenger Ribonucleic Acid g microgram l microliter ml milliliter min minute NPV Negative Predicted Value NSMB Normal Suckling Mouse Brain PPV Positive Predicted Value PRNT Serum Neutralization Plaque Reduction Tes t SLEV St. Louis Encephalitis Virus VEEV Venezuelan Equine Encephalitis virus WEEV Western Equine Encephalitis virus WNV West Nile Virus


viii Luminex Microsphere Immunoassay Offers an Improved Method in Testing for Antibodies to Eastern Equine Encephalitis Virus in Sentinel Chickens Kelly Ann Fitzpatrick ABSTRACT Eastern Equine Encephalitis virus has a human morta lity rate of 30% of those cases diagnosed, while 30% of those surviving infec tion remain with neurological sequelae for life (, 2007). The use of sentinel chickens for surveillance of a rboviruses that are known to use birds as a reservoir host, such as St. Louis Enceph alitis (SLE), West Nile (WN) virus, Eastern Equine Encephalitis (EEE) and Highlands J ( HJ) virus, in Florida began with the Sentinel Chicken Arboviral Surveillance Network in 1978 (Day and Stark, 1996). This network enables the activation of an early warning system for citizens, as well as, county epidemiologists and those in mosquito control, allo wing for a coordinated effort of disease prevention. Methods currently used at the Florida Department o f Health, Tampa Branch Laboratory include screening of submitted sera for antibodies to these arboviruses of epidemiologic importance by way of the hemagglutina tion inhibition test (HAI), and confirmation by the IgM antibody capture enzyme-lin ked immunosorbent assay (MACELISA) and Plaque Reduction Neutralization test if the MAC-ELISA proves to be


ix negative. While these tests combined are providing the results needed, the time to result can be a week or greater depending on the initial s creening result in the HAI tests. The Microsphere Assay Technology provides an accur ate, more rapid (a day or two instead of a week or more) detection method inc luding both a screening and confirmation protocol specifically designed to test for antibody to EEE in sentinel chicken sera. Two sera out of the thousands tested that were tested by HAI shown to be negative in standard testing, resulted as positive by the MIA method and therefore indicated a missed positive. The sensitivity and sp ecificity, positive and negative predictive values of this new protocol as compared with MAC-ELISA as a reference standard indicated that both tests were remarkably similar; Providing sensitivity near 80%, specificity and PPV at 99%, and negative predi ctive values at 90% for MACELISA and 94% for the MIA. Finally it was determine d that Highlands J virus will not have any impact on the testing protocol and results of this test.


1 Introduction Arthropod-Borne Viruses Requirement of an arthropod transmission vector is a unique characteristic of arthropod-borne viruses (arboviruses). The endemic ity of these viruses is dependent on three main components, including the virus, vector and the vertebrate host. These three components dictate the spread, severity and impact upon humans by each virus. Environmental factors are also an important impacti ng and controlling aspect of vector borne diseases; they have distinct effects upon the lifecycle of the known vectors thereby directly affecting the transmissibility of the viru s to new hosts. These factors include temperature and rainfall, thereby directly affectin g the transmissibility of the virus to new hosts. Vectors within the Phylum Arthropoda include mosquitoes, biting flies and midges, sand and black flies, mites and ticks, all capable of pathogen transmission to a suitable vertebrate host. In general, vectors have the ability to transmit both mechanically as well as biologically, the latter being vital to the propagative transmission of arboviruses; thus, arboviruses depend upon their ve ctor hosts for multiplication as well as transportation between hosts (Chamberlain 1961). Currently, the focus of public health in the Unit ed States is concentrated only on a short list of arboviruses including members of the Flaviviridae, Bunyaviridae and Togaviridae families; these include West Nile virus (WNV), Lacrosse virus (LAC), St.


2 Louis encephalitis (SLE), Western Equine Encephalom yelitis (WEE), and Eastern Equine Encephalomyelitis (EEE) viruses. Species of hematop hagous mosquitoes are the known primary vectors of these viruses, infecting fowl wi th a cyclic, mostly enzootic pattern. Mammals including humans and horses are also involv ed in many cases as dead end hosts, which are defined as those hosts unable to p rovide sufficient viremia to further infect subsequent potential vectors or hosts. Thr ough the work of Cupp et al. it is believed that ectothermic species, such as snakes a nd other amphibians could potentially be an overwintering reservoir host in the southeast ern United States for Eastern Equine Encephalitis (Cupp et al, 2004). The use of sentinel chickens for surveillance of a rboviruses that are known to use birds as a reservoir host, such as St. Louis Enceph alitis (SLE), West Nile (WN) virus, Eastern Equine Encephalitis (EEE) and Highlands J ( HJ) virus, in Florida began with the Sentinel Chicken Arboviral Surveillance Network in 1978 (Day and Stark, 1996). Sentinel chickens are young immunologically nave c hickens that have not previously been exposed to the viruses being assessed and are therefore sero-negative, or do not have antibody specific to viral proteins in their s era. This network enables the activation of an early warning system for citizens as well as county epidemiologists and those in mosquito control, allowing for a coordinated effort of disease prevention. Prevention includes control of the vectors, as well as warning for medical personnel of possible encephalitic disease in the area, and the potential need for treatment of those individuals. Warnings for the public include press releases and media advisories, which are critical steps taken to inform the public to take defensive action against being bit by mosquito vectors.


3 There are 550 arboviruses currently identified, at least 100 of which are known to be pathogenic to humans (Gould 2006, Cann 2001). Th ese are grouped into four viral families, which are the Togaviridae, Flaviviridae, Bunyaviridae and Arenaviridae (Table 1). All arboviruses are enveloped, measure 17-150 nm or more and are often spherical or occasionally rod shaped. Their genomes are coded as either positive or negative sense, single stranded RNA genomes. These RNA genomes are known to experience frequent mutation, thereby increasing the potential genetic variation within species and their phenotypic and genotypic differences (Calisher, 199 4). These viruses infect humans by way of passage through the blood brain barrier from the peripheral circulation, where initial inoculation took place. Arboviruses have evolved pathogen-vector relations hips enabling not only their multiplication but also transmission from one host or species to another, thereby ensuring their ability to survive. In fact, scientific specu lation suggests that ticks may have actually been the vectors first involved in the evolution of certain viral lineages, followed subsequently by the transmission by mosquitoes (Kun o and Chang, 2005). It also was once believed that viral multiplication was done wi th little to no damage caused to the arthropod host; recent studies however have found c ertain pathologies and decreases in function may occur to the vector during some infect ion with some viruses (Kuno and Chang, 2005). In order for infection by the virus to take place, the vector must first be susceptible to infection and second, ingest a minim um threshold level of virus from the primary source (Chamberlain and Sudia, 1961). It is believed that the ability to be a


4 Table 1. Arthropod-borne viral families of signific ance: major Arboviruses and their significant non-arbovirus relatives (*) and the loc ations where they can be found. (ICTVdB Management, 2006) nr nrnrn nrnrrn nr nrnnrnr nrrnnr nr!rnnr rr "#nnrnnnr"#n nr $rnrn nr r %rn&nnnr '(rnrrn nr $rr&nnnr ))rnrn *)rnn +nnnr,+rnrr nrr,$) +nn nnnr r n r,nnr nnr /"nrnnnn* nn r .nr nn (nrrnr,$rn


5 suitable vector is dependent upon the presence of c ertain binding receptors in the gut of the vector (Kuno and Chang, 2005). While evidence f or this appears to be strong, primary studies have only been conducted using data extrapo lated from the use of cell culture not studied in vivo, and therefore conclusions must be taken carefully if mosquito derived cell cultures were used (Kuno and Chang, 2005). The minimum infectious dose of the arboviruses varies by virus and species of vector; however, a correlation has been drawn between the threshold levels for arthropod infectio n and viremias in selected vertebrate hosts (Chamberlain and Sudia, 1961). Once infected, the vector, after an extrinsic incubation period, remains infected for the remaind er of its life, for mosquitoes this is measured in terms of days or weeks, ticks on the ot her hand, due to their ability to become infected in their immature stages, may be in fectious most of their two year life span (Kuno and Chang, 2005). Primarily arboviruses are known to be zoonotic dis eases, passing from one animal host to another via the vector intermediary. Suscep tibility to infection is a primary determinant for the successful transmission and dis semination of arboviruses, both in the vector as well as the vertebrate host. Therefore, d evelopment of herd immunity or a decrease in population of the reservoir host has a significant impact on the continuance of infection transmission (Kuno and Chang, 2005). Arbo viruses have developed three main strategies to ensure survival when these two limiti ng factors are in place. They include exploiting the inherent mobility of their vectors t o new host populations, selecting for host vertebrate species with a high fecundity rate so that new nave hosts are readily available, and lastly incorporating the means to ev ade the immune response in their hosts (Kuno and Chang, 2005).


6 Western and Eastern Equine Encephalitis viruses are both categorized as alphaviruses or Group A viruses, by terminology onc e used by the Center for Disease Control and Prevention (Kissling 1960); as of today there are known to be 27 alphavirus members identified (Calisher, 1994, Baron, 1996). W estern equine encephalitis (WEE) is principally vectored by Culex tarsalis between passerine birds and intermittently a mammalian such as horses or humans (Kissling, 1960, Passler and Pfeffer, 2003)). It can be found at its peak between the months of June and September in western North America, in South America and Cuba however its tran smission cycles are only narrowly understood (Kissling, 1960, Weaver et al, 1997). Th e extrinsic incubation time or time of multiplication within the mosquito has been determi ned to be approximately 8 days but may become infectious in as little as four days (Ki ssling, 1960). 639 confirmed human cases of this virus have been known to occur in the United States since 1964 (, 2007). Also a member of the alphavirus genera is Eastern E quine encephalitis (EEE), while this virus will be discussed in more detail l ater; it is known that the primary enzootic vector in the infection of birds is Culiseta melanura (Kissling, 1960)(Service, 2004). The primary vectors to the mammalian hosts include members of Aedes, Anopheles, Culex, Ochlerotatus and Culicoides (a biting midge) species, as well as a the bridging vector between peridomestic and sylvatic c ycles Coquillettidia perturbans (Kissling, 1960)(Service, 2004). Peak transmission can be found between March and September in the United States (Kissling, 1960). Th ere are four lineages of EEE virus Group I occurring in North America and the Caribbea n with the greatest health impact and primarily equine related in Central and South A mericas caused by IIA, IIB, and III


7 (CDC, 2005). On average, 5 cases of human disease o ccur per year; between 1964 and 2004 approximately 220 human cases were confirmed (, 2007). EEE has a mortality rate of 50-75% of diagnosed cases, while 30% of those surviving infection remain with a neurological sequelae for life (Chonm aitree et al., 1989) West Nile virus (WNV) and St. Louis Encephalitis vi rus (SLE) are both members of the Flavivirus genera also known by older CDC no menclature as Group B viruses; they encompass 27 distinct viruses, including dengu e, Japanese B and yellow fever. Culex species including Cx. pipiens, Cx. modestus and Cx. univattatus are all known vectors of West Nile virus for humans as well as bi rds, which may also be infected by certain ticks according to Service (2004). From 196 4-2006 there were 27,573 cases of WNV in the U.S. reported to CDC, in 2006 among pati ents 33% had encephalitis symptoms, 65% had a milder form referred to as West Nile Fever and 59% had unspecified symptoms, SLE however from 1964-2006 ha d only 4,658 cases reported, fatalities among cases of SLE are generally approxi mately 5% of the infected, while those with life long neurological damage may includ e 10% of those surviving ( DVBID, 2007).


8 Alphaviruses Epidemiology and pathogenesis Alphaviruses and Rubiviruses are the two genera tha t make up the family Togaviridae (Schmaljohn and McClain, 1996, Powers et al., 2001 ). Alphaviruses are present on all continents of the Earth with the exc eption of Antarctica (Powers, 2001). Due to their individual ecological cycles, amplifyi ng and reservoir hosts, as well as their often highly specific hematophagous vectors, they r emain relatively focused in their ideal niches such as swamps or woodlands, western or east ern North American geographies (Powers, 2001). There are seven acknowledged seroco mplexes, or antigenic types of alphaviruses, three of these are of medical importa nce; they include Eastern Equine Encephalitis viruses (EEE) subdivided into North an d South American, Western Equine Encephalitis viruses (WEE), and Venezuelan Equine E ncephalitis viruses (VEE) (Passler and Pfeffer, 2003, Schmaljohn and McClain, 1996). EEE can be found predominantly in a natural cycle b etween songbirds in fresh water swamps. Research by Unnasch et al. indicates that the cyclic behavior exhibited by this virus may be due to feeding by mosquitoes on n ave young of year (YOY) birds, mainly by the ornithophilic mosquitoes like Culiset a melanura, this research focused on the middle to eastern parts of the United States in cluding almost the entire Atlantic coast from Florida to New Hampshire (Schmaljohn and McCla in, 1996, Weaver et al., 1994, Unnasch et al., 2006) WEE, isolated from birds, horses, humans and other mammals can be found in both Canada and the western United Stat es (Calisher, 1994). VEE unlike EEE and WEE does not appear to use birds as a reservoir host; it is believed that the cycle


9 includes a mosquito to small mammal and back to mos quito pathway (Schmaljohn and McClain, 1996). The last major epizootic of this d isease was seen from 1969 to 1972 whereby it entered the United States from Mexico th rough Texas (Schmaljohn and McClain, 1996). Alphaviruses in general produce a variety of disea se processes and symptoms; these can include a fever, macular-papula r rash, arthralgia, malaise, and arthritis in Old World viruses such as Ross River, Barmah Forest, Mayaro, Chickungunya and Sindbis, and encephalitides in the New World vi ruses such as EEE and WEE (Powers et al., 2001). They can infect a variety of creatures including rodents, reptiles, amphibians, fish, horses, humans and birds through the passage and multiplication within their arthropod vector with the exception of the sa lmonid viruses which include Salmon Pancreas Disease virus and Sleeping Disease virus i solated only from salmon and rainbow trout and not currently from any vector spe cies (Powers et al., 2001). Each of these species have the ability to mount a specific immune response to the infective agent by way of the proteins protruding through the host derived phospholipid bilayer present surrounding the viral capsid protein. A specific im mune response provides an opportunity to not only assess contact with the agent but in so me cases the approximate time frame of contact. (Schmaljohn and McClain, 1996) Classification and antigenic types Classification of the alphaviruses has generally been based on the antigenic relationship of the members (Schmaljohn and McClain 1996). The amino acid


10 sequences of the envelope glycoproteins E1 and E2 f ollowing PCR amplification is a more recent method used to identify and understand the evolutionary relationships of the alphaviruses (Bell et al., 1984, Schmaljohn and McC lain, 1996). This technique originates from the original grouping that was base d upon serological cross reactivity in hemagglutination-inhibition testing, complement fix ation and the plaque reduction neutralization tests (Bell et al., 1984). Immunological Response Human immunity consists of two phases of response to an invading agent. These are the innate immunity for the early reaction and the adaptive or specific immunity for the later response. Of most importance for serologi cal testing methods of acute disease is the response of the later phase, the adaptive respo nse; the adaptive response itself is divided into two types, cell mediated and humoral i mmunities. Cell mediated immunity involves T lymphocytes while humoral immunity conce rns the antibodies made by B cells. Antibodies contain binding sites specific to each invaders surface protein or other antigenic components. These immunoglobulins have th e ability to target an invader for phagocytosis and destruction, neutralize its infect ious ability, limit the toxicity of toxins and prevent entry of viruses into host cells. There are different types of antibodies produced by B cells, these are different depending on the B cells state of activity, whether the cell is still nave or has been activated by T lymphocytes and contact with a recog nized antigen. IgM is a pentamer molecule and is the first immunoglobulin (Ig) molec ule secreted from B cells while it is still a nave cell, prior to isotype switching to I gG. IgM is a primary antibody response,


11 taking place when an antigen makes contact for the first time. After the immune response to the antigen is complete and the infection subsid es these B cells will become memory cells and are kept for a period of time in case the y are needed again. IgG is a secondary response, this Ig begins to be produced after conti nuous activation by the antigen and T lymphocytes. As the initial infection proceeds the cells begin to switch their specificities, IgG then becomes the prominent class over IgM in th e case of most bacteria and viruses. IgG is a monovalent Ig and in humans has several di fferent heavy chain isotypes. (Pier, 2004)(Janeway, 2005) Because IgM is the primary antibody response to vi ral infection, quantification of it is an indicator of a new response to that infect ion, while IgG response alone simply indicates that at some point the individual has rea cted to this invader before. IgM can also be seen as a subsequent response to the same (or in some cases cross-reactive) antigens, it is however usually in lower quantities than an init ial response. In non-vertebrate mammals, such as the chicken, Ig isotypes vary slightly from vertebrate mammals. Avian IgY is a homologue to hum an IgG with some slight variation; IgY contains four CH domains while human IgG contains only three, it st ill however displays the same functional properties (Viertlboec k, 2007). Chickens also produce IgM and IgA, IgM being produced prior to IgG (IgY) as i t is in all vertebrate species (Johnson, 2003). It is this early Ig that is the focus of the MAC-ELISA and the current Luminex protocols employed by the Florida Department of Hea lth.


12 Eastern Equine Encephalitis virus History Eastern Equine Encephalitis (EEE) of the genus Alphavirus and the family Togaviridae appears to run in both sylvatic and peridomestic c ycles (Figure 1) (Kissling, 1960). This virus is highly focalized due to ecolog ical factors including the seasons, weather, the host/vector immunity and the density o f populations. It was first isolated in 1933 from the brain of a horse, though the possibil ity of its presence in North America goes back to 1831 when a disease of similar clinica l symptoms was documented after the death of 75 horses due to an encephalitic disease, though not confirmed (Calisher, 1994, Evans, 1977). The first human cases were confirmed in 1938 after the death of thirty children in the northeastern region of the United S tates following an epidemic in horses (Calisher, 1994). Culiseta Melanura is the vector responsible for the cyclic transmiss ion among swampy and fresh and salt water marsh habitat s (Calisher, 1994). This vector feeds primarily upon the wild bird populations in t hese areas and permits enzootics and epizootics that have the ability to move outward fr om the permanent foci of the virus, among the marshland wild birds (Calisher, 1994, Kis sling, 1960). The intersection of the epidemic vectors into the scheme also takes place w ithin the swampy marshlands. The epidemic vectors Aedes sollicitans and Mansonia perturbans were initially the main epidemic vectors in the United States; they are slo wly being replaced however since the 1985 introduction of Aedes albopictus (Calisher, 1994).


13 Figure 1. Diagram of the sylvatic and peridomestic cycles of Eastern Equine encephalitis. Wild Birds in swampy woodlands and marshes C. melanura Culex, Aedes, Coquillettidia Dead End Hosts Humans, Horses, Domestic Birds, Reptiles, other mammals Enzootic cycle Peridomestic cycle


14 The epidemic vectors are responsible for the trans mission of the infection from wild birds in the marshes to human and animals outs ide this ecological niche including dead end hosts such as humans, horses, swine, deer, small mammals, and reptiles. These dead end hosts, while having the ability to become infected by the virus, rarely establish a viremia significant enough to permit transmission t o nave mosquito vectors; although it has been documented that horse to horse transmissio n can take place through A. sollicitans if the horse’s viremic titers are found to be abov e normal (Kissling, 1960). Species of mosquito that are known to become infect ed and transmit EEE have been shown to require different viral titers for infecti on. Certain Aedes and Psorophora sp. have been shown to require a viral titer of are lea st 103.0 LD50 per ml, several species of Culex have a requirement of at least 108.0 LD50 though C. tarsalis needs only 102.5 LD50 per ml (Kissling, 1960). The percent efficiency of infection is directly related to the viremic titer and therefore changes differently bas ed upon vector and host species. Epidemiology Eastern Equine Encephalitis virus can be subdivide d into the North American, (also including the isolates from the Caribbean) an d South American (also including isolates from Central America) subtypes (Calister, 1994, Passler and Pfeffer, 2003, Weaver et al., 1994). These two subtypes differ fro m each other in epidemiological, biological and genetically characteristic ways, the reby allowing isolation methods such as complement fixation (CF), hemagglutination-inhibiti on (HAI), enzyme-linked immunosorbent assay (ELISA), plaque reduction neutr alization test and nucleotide


15 sequencing for distinction of the pathogen for rese arch and diagnostic testing (Passler and Pfeffer, 2003, Strizki and Repik, 1994). In fact, i t was Casals et al. who were able to show homogeneous reactions between virus located in North America and the Caribbean using kinetic HAI tests (Calisher, 1994). The South American subtype appears to cause fatal infection in horses; however clinical symptoms in the rare human cases appear si gnificantly less severe than its North American counterpart (Calisher, 1994, Passler and P feffer, 2003). Infection by the North American subtype while also rare in humans, compare d to infection by some of those in the Flavivirus group, appears to have severe clinic al manifestations for both horses and humans (Calisher, 1994, Passler and Pfeffer, 2003). The North American subtype is annually found in horses along the eastern coast, a s far north as the southeastern portion of Canada and as far west as the upper Midwest Unit ed States (Calisher, 2003). The continued study of West Nile virus has brough t about another possible mode of transmission, organ and blood transplantation. The first report of WNV transmission in the blood supply occurred in 2002 (CDC, 2007). This then prompted the nationwide screening of the blood supply using minipool nuclei c acid-amplification testing (MPNAT) (CDC, 2007). Due to the extremely rare nature of the EEE virus in humans it is not likely that this mode of transmission will be of co nsiderable significance. However it is important to keep vigil of the country’s much neede d blood and organ donation supply especially considering that many requiring its use may have immune compromise.


16 Host Vector and Transmission While arboviruses are spread by a variety of arthr opods, Eastern Equine encephalitis is primarily transmitted by mosquitoes There are approximately 3300 species of mosquito in the world, all of which are contained within the family Culicidae (Service 2004). Three subfamilies that branch from this include the Culicinae, Anopholinae and the Toxorhynchitinae. EEE is prima rily transmitted by members of the Culicinae subfamily, although Kissling and also Wel lings both reported isolation of EEE virus from Anopheles crucians (Kissling, 1960, Wellings et al, 1972 ). In the sy lvatic cycle the ornithophagic mosquito Culiseta melanura transmits the virus during the acquisition of a blood meal. According to research by Unnasch et. al nestlings and youngof-the-year (YOY) passerine birds provide a large m ajority of the meals for this endemic vector. These nave young birds tend to form circul ating viremia faster, and to levels equal or greater than those found in adult birds (M cLean et al. 1995, Unnasch, 2006). In the peridomestic cycle, that which interacts with t he birds, as well as humans and in limited areas horses, the primary bridging vectors includes several species of Oclerotatus, Aedes, Culex and Coquillettidia, again all members of the subfamily Culicinae (Kissl ing, 1960, Service, 2004). As all mosquitoes, those in the subfamily Culicina e have a life cycle involving larvae, pupae and then adult phases. The eggs in th is family are normally dark in color (brown or black) but can be laid either singly or i n rafts depending on the genus (Service, 2004). All mosquitoes require water for egg deposit in some form, Culiseta melanura typically prefers fresh water, such as swamps or la kes, Culex prefers ground collections


17 of water as apposed to Aedes and Oclerotatus which prefer smaller container habitats which include such things as tree-holes, tires, poc k pools and plant axils (Service, 2004, Davis, 2005). Larvae also have distinguishing chara cteristics, Culex, Oclerotatus and Aedes all breath air directly from a siphon at the surfa ce of the water in which they reside. Coquillettidia on the other hand, similar to the Mansonia genus, have siphons that provide them the ability to insert it into the root s of plants that float on the surface (Service, 2004). This has a significant impact on the methods of mosquito control for each type, those that reside on the surface can be reached by pesticide much easier than those submerged, however, those dependent on plants for air might alternatively be indirectly controlled by controlling the plant grow th (Service, 2004). The pupae of each genus remains in the habitat of deposition, their b reathing characteristics remain the same as during their larval stages (Service, 2004). Upo n emerging from the pupal casing the adult mosquitoes vary widely in physical characteri stics including color and pattern of their scales on wings, abdomen and legs. These very distinctive differences in appearance allow entomologists and those in mosquito control t o identify the many varieties. While it was long believed that infection of mosqu itoes with an arbovirus caused no deleterious effects upon its host, it is now und erstood that infection with the virus may have an effects upon the hosts’ fecundity, survival and obtainment of a blood meal (Moncayo, 2000). Moncayo et al. discovered that di sseminated EEE infection had a significant impact on the length of survival (7-14 days) of Cq. perturbans and Cs. Melanura in their 20 day per os experiment though there was little effect seen in An. Quadrimaculatus or Ae. Albopictus, which lived throughout the time of the experiment (Moncayo, 2000). The survival of the mosquito post extrinsic incubation period is vital in


18 determining vectorial capacity and therefore the ti me the virus can be transmitted through the vector as well as its rate of replication (Monc ayo, 2000). Clinical Features of EEE Humans There are an average of five cases of EEE each yea r in the United States, though this fluctuates by the year from 0-15. Florida, Geo rgia, Massachusetts and New Jersey have some of the highest case numbers seen by the C DC. While Human cases are very rare the sequelae post infection in many of those i nfected can be permanent and require the individual to be placed in permanent institutio nal care (Kelso, 1999) or lifelong disease related expenses nearing 3 million dollars per patient (CDC, 2005). Symptoms of EEE are usually seen within three to ten days of be ing bitten by an infected vector, it has an infection rate of 33% and if infected 50-70% of cases end in mortality and only 10% fully recover with no morbidity (Nandalur, 2007, CD These statistics are based upon those who have had sera submitted for testing for viral antibodies; these numbers have not been fully assessed in a surveillance test ing of the general population. These symptoms include: sudden onset of fever, vomiting, leukocytosis, hematuria, stiff neck, headache, malaise, and general muscle pain, and may progress to meningitis and /or encephalitis of the brain, seizures coma or paralys is. By the time symptoms are present virus is no longer found in the blood (Chonmaitree, 1989) therefore public health laboratories are usually called upon to test for an tibodies to the virus in blood or cerebral


19 spinal fluid to determine if infection is recent (I gM) or from a previous period (IgG). Testing for these antibodies produced in the later phase of infection includes such techniques as hemagglutination inhibition (HAI) tes ting, MAC-ELISA, serum neutralization (SN) (Calisher, 1986) and complement fixation testing (CF). Confirmation of damage by EEE infection can be done with magneti c resonance imaging and computed axial tomography or CAT scan, provide evidence of e dema, ischemia, and hypoperfusion in the early stage and proceeding to necrosis, vasc ular hemorrhage in brain and visceral organs and encephalitis (Paessler, 2004). Horses and other ungulates Horses are also a dead end host for EEE. Mortality in horses is between 75-90%. As in humans the disease can cause permanent neurol ogical damage. In 2003 a record number of horses were infected with EEE, cases have risen nearly 3,000% in some parts of the United States. Clinical symptoms include a l ack of coordination, loss of appetite, the grinding of teeth, inability to swallow, circli ng, involuntary muscle movements, rear end weakness, blindness, excitability and sensitivi ty to light and sound (, 2003). These symptoms can then progress to further in-coordination, hyperesthesia, paralysis, coma and seizures which occur usually wi thin 48-72 hours of clinical signs, eventually leading to death (, 2003, Davis, 2005). Several clinical signs of EEE may be misinterpreted as rabies or toxicosis or vice versa, these include aimless wandering, circling, head pressing, staggering gait and difficulty swallowing, viral isolation or antibody testing can be used to differ entiate the diagnosis from other


20 potential diseases such as rabies or some physical disease process such a stroke. MACELISA is commonly used to test horse sera; a follow up paired sera for confirmation is also usually submitted to confirm diagnosis by PRNT PCR is used to test for viral titer also in blood, though viremia is commonly transient although commonly diagnosis is made upon necropsy (Poonacha, K.B., 1998) Recovery takes weeks to months of gradual improvement, though they may never fully recover pr e-infection abilities (Davis, 2005). Those horses that do survive disease are often refe rred to as “dummy” horses due to the long term damage sustained to their brain, these ho rses can not be safely used to breed or ride (Jacob, 2003). During the early phase of disease, when viremia is at its peak, is the most effective time to implement what little treatment i s available for horses (Davis, 2005). This includes hyperimmune plasma or serum products infused immediately because by the time the neurological symptoms begin the viremi c phase is most likely to be past (Davis, 2005). Truly vaccination is the best course of action. Vaccination for EEE, WEE and VEE in adult horses should be done bi-annually starting in early spring and early summer. In areas that experience endemicity however the horses should receive boosters every six months (Davis, 2005). It has been shown t hat vaccination IgM should not interfere with the founding of the diagnosis, shoul d the need arise; in other words the IgM level should never reach a four-fold rise in titer (Davis, 2005). Pathological changes in the infected horse may inc lude “neuronal degeneration and necrosis, vasculitis, and vascular thrombosis, gliosis, and neutrophilic infiltration” of the central nervous system including the cerebral c ortex, thalamus, hypothalamus, and the anterior portions of the spinal cord (Poonacha, 199 8, Franklin, 2002). One case report


21 indicated a colt with smooth muscle necrosis of the tunica muscularis of the small intestine post vaccination with a killed EEE vaccin e, though the case is considered to be naturally occurring (Poonacha, 1998). Another case report indicated involvement of the bladder tissues when hemorrhagic tissue was found d uring necropsy also following 1 week post-vaccination (Franklin, 2002) Though thes e are the only mention found of such involvement outside the nervous system there is als o a known case of myocardial involvement post infection in a pig (Poonacha, 1998 ). Pigs, cattle, and goats exhibit similar symptoms a s horses, however convulsions and paddling are added as more severe signs of path ology (Farrar, 2005). It is important to keep surveillance of this infection in horses an d other barnyard animals in mind, due to their proximity to humans; disease in those unvacci nated indicates vectors present in the area are infected with the virus. Also important i s the potential economic impact of the disease, racehorses, found commonly in Florida and other southern states cost an average of $95,000 to purchase and $35,000 dollars a year t o train, this of course does not include those equines that are pets and still may be a cons iderable financial and personal loss for their owners. Another important consideration is th e economic impact this disease may have upon agriculture. While human cases may be few animal cases are known to be more common, this may impact food production in tho se animals raised for consumption such as pigs and cattle. Another consideration is t he effect of infection in the long term for those animals that survive, is it possible that the animals or the products produced from surviving animals contains anything harmful to other animals of the industry or to the humans that consume them?


22 In Houston County Georgia in 2001 Tate et al. diag nosed EEE in a male white tailed free ranging deer ( Odocoileus virginianus ). Upon necropsy the deer was found to have histological finding similar to those found in EEE infected horses (Tate, 2005). In this same region the death and finding of the deer was following the death of two horses of the same virus (Tate, 2005). Antibody studies of deer in the region followed, 32% of the deer sampled were found to have antibodies to E EE (Tate, 2005). In one area the percentage reached as high as 55%, this indicates t hat deer are certainly exposed to the virus, however because there is little knowledge of symptoms or reports of outright disease they are probably not susceptible to severe infection (Tate, 2005). Other Mammals Other mammals documented to be susceptible to infe ction by EEE virus include, rodents, dogs, and bats. Studies in mice indicate that the virus spreads and multiplies through the peripheral tissues, into the blood stre am and from there infects the CNS and brain tissue (Vogel, 2005). The symptoms found in m ice include lethargy and ruffled fur initially, then after 4 days post-infection they ex hibit to hunching, tremors, and prostration, which may then progress to death (Voge l, 2005). Hamsters and guinea pigs are also susceptible to infection by EEE and althou gh it is not a mammal, inoculation of the spotted turtle ( Clemmys guttata ) has been shown to develop viremia (Smith A.L., 1980). The susceptibility of these small creatures of EEE has provided the opportunity to study vaccines in hamsters such as the work done by Cole and McKinney on a trivalent vaccine for EEE, VEE, and WEE (1971). Signs of infe ction in the hamsters initially


23 include pressing of their head against the cage, vo miting, lethargy, and anorexia; proceeding to central nervous system involvement af ter four days infection including stupor and coma then death (Paessler, 2004). These small mammals provide scientists with models in which to study the infectious diseas e process of EEE in an animal model with similar immunological response to humans it al so allows the in-depth study of the pathogenesis of disease in humans in an economicall y feasible way. In a twelve year span a veterinary clinic in South Georgia saw the deaths of 101 cases of dogs with neurological disturbances (Farra r, 2005). Symptoms included pyrexia, anorexia, and diarrhea, followed later by recumbanc y, meningitis, nystagmus, depression and seizures (Farrar, 2005). The clinic notes no sp ecific breed that was over represented, however all dogs were known to be housed outside pr imarily (Farrar, 2005). Upon necropsy the dogs were seen to have infiltrates in the gray matter of the brain predominantly in both their cerebral cortex and the ir midbrains; each diagnosis was confirmed by positive viral isolation by either tis sue or blood samples (Farrar, 2005). Bats are yet another mammal that has been shown to become infected by EEE. Colonial bats from Massachusetts, Connecticut, New Jersey and Georgia, are shown to be naturally infected by mosquitoes (Main, 1979). Acco rding to the work by Andrew Main at Yale University, it is believe the virus is spre ad to the host via the bite of the mosquito and not ingestion of the vector (Main, 1979). CNS i nvolvement in bats is rare and therefore encephalitic symptoms are also rare, this may be due to the lack of neurotropism seen in Main’s study of hibernating co lonial bats (Main, 1979). While it has been thought that bats have the ability to be overw intering hosts, Main was able to detect viremia from hibernating bats only 24hr to 42 days from bats collected from caves and


24 mines in Massachusetts, Connecticut, New Jersey and Georgia (Main, 1979). These bats were bled and organ samples obtained following one week in an environmentally controlled cabinet in the lab (Main, 1979). Potent ial methods of transmission among the bat population include transplacental, through moth er milk or possibly through urine (Main, 1979).. Another important finding of this re search was the limited number of bats positive for EEE in their saliva, thereby making it unlikely that transmission could proceed by the bite of these animals unlike rabies (Main, 1979). Fow l Birds are the known reservoir hosts for EEE. All b irds including those wild, as well as raised domestically are susceptible to infe ction, domestically reared fowl such as emus can have an infection rate of 65 percent with a mortality of 80 percent but can be as high as 100 percent, pheasants have a potential mor tality of 25%-100% of those infected, turkeys considerably less at 5% mortality and ducks up to 60% can be lost due to infection with EEE (Helm, 2003). Signs of infection depend upon the species, but may generally include depression, bloody diarrhea, vomi t, drowsiness, in-coordination, anorexia, blindness, partial paralysis and neurolog ical involvement (Jacob, 2003, Helm, 2003). Emus in particular may become infected with out vector involvement; their bloody feces can contain such high levels of viremia that it could be infectious to both other birds as well as humans (Helm, 2003). Supportive care is really the only treatment for infected birds, attempting to limit damage to the central ne rvous system and secondary infections.


25 Chickens tend to have greater viremia during infec tion when they are younger before the 14th day of life, a period prior to use by a sentinel c hicken program (Guy, 2003, Byrne, 1960). The lethality of the infection decreases as the chicken ages (Guy, 2003). Signs in young chickens include depression, somnolence, paralysis, and death, though these are rarely seen (Guy, 2003). Myocardit is is attributed to be the most significant cause of death in chickens post EEE dis ease, though microscopic lesions are seen occasionally (Guy, 2003). Other findings on ne cropsy include necrosis of the liver, thymus, spleen, and bursa of Fabricius (Guy, 2003). Chickens are commonly used as sentinels for Arbovirus surveillance programs, the older the bird the less likely their infections will reach viremic levels sufficient to contribute to amplification in their immediate environment; they are relatively easy and inexpensive to maintain and bleed, and can be maintained relatively easily in populate d areas where infection concerns may be greater than extremely rural populations (Moore et al. 1993). Treatment and Prevention Though various testing methods are now available fo r diagnosis of this viral disease, there are few if any treatments available to those infected. Pharmacologic supportive therapy for humans includes, antipyretic s, analgesics, and anticonvulsants in addition to physical intervention such as ventilato r support if patient becomes comatose (Nandalur, M, 2007). Although it is currently only in the research phase Ribavirin, pyrimidine derivatives and isoprinosine have been a ssessed for any attenuation of the infection in vitro, though in vivo results have bee n questionable (Nandalur, M, 2007)


26 Prevention is the best way to limit the spread of this disease. To do this people must be aware that mosquitoes are the vectors of tr ansmission and by preventing bites from them, they are in fact protecting their health Many of the Mosquito Control agencies of Florida and the Florida Department of H ealth all stress the five D’s of prevention ( tml, ( m) (Heshmati, 2004) and the one S These include avoiding being unprotected outside at dusk and dawn, the peak biting times of mosquitoes, dressing with clothing that provide coverage of the skin and therefore less skin accessible to biting, donning repellent with D EET, and draining of containers with standing water. The one S added by the Florida Depa rtment of Health includes the use of screens of windows to prevent mosquito entry into t he home (Heshmati, 2004). As previously mentioned many of the genus’s within Cul icinae are container breeders, therefore removal or draining of a potential breedi ng container will decrease their numbers in the immediate vicinity of the container such as a residence if there is an open rain barrel. Arbovirus Surveillance Surveillance for arboviruses varies depending upon the state. Programs can include counting or viral detection in vectors, ser ological testing on sentinel or wild vertebrate hosts and case detection among domestica ted animals such as horses as well as humans. A good surveillance system takes into accou nt seasonal dynamics, ecology, meteorological data, and vector and vertebrate host surveillance. It is this last category


27 that is focused upon by the Florida Sentinel Chicke n Program. According to the CDC Guidelines for Arbovirus Surveillance Programs in t he United States by Moore et al. (1993), arbovirus vertebrate hosts for surveillance should have the following characteristics: “1. Susceptibility to the monitored virus at rates that reflect virus activity in the surveillance area, 2. High Titer and long duration of antibody respon se, 3. Low morbidity and mortality (except in those sp ecies where high mortality is easy to detect), 4. Locally abundant population, 5. Locally mobile to increase exposure to and diss emination of virus, 6. Frequent exposure to vector species (could over come lack of mobility), 7. Attractive to and tolerant of vector feeding, 8. Easily captured by conventional methods, 9. Ease in handling and obtaining blood specimens, 10. Age determination possible, at least young of year, or the regular multiple captures of tagged animals permits detection of seroconversions, 11. Relatively long-lived for multiple sampling of same animal”. With this information in mind there are several ava ilable surveillance hosts from which to choose. These include chickens, wild birds equines, domestic and wild mammals, mosquito and lastly human case surveillanc e. Each meets the criteria listed above though each has its challenges. Chicke ns are relatively inexpensive to maintain and do not require large amounts of space, they must simply be placed


28 strategically so that infection can take place in a reas commonly inhabited by people (Moore et al, 1993). Chickens can be used for WN, SLE, WEE, or EEE viruses though they have not been found useful in New Jerse y for EEE (Moore et al, 1993). Use of equines is usually either a passive or rarel y an active surveillance method. Due to the expense of raising and keeping horses it wou ld not be economic to raise horses purely for the use of surveillance. Cases can howev er be quantified using passive surveillance of veterinarians. However in the cases of horses their vaccination status, transportation and delayed veterinary reporting may affect the data received and thereby effect the conclusions and activation of th e warning system (Moore et al, 1993). Florida Sentinel Chicken Program The Florida Department of Health and Rehabilitative Services began the Florida Sentinel Chicken Program Arbovirus Surveillance Pro gram in 1978 during a rural epidemic of SLE virus (Day and Stark, 1996). This s ystem provides detection and early warning for areas, like Florida, where arboviruses at times become epizootic and spill over into the human populations causing epidemics. This system currently focuses primarily on West Nile virus (WNV), St. Louis Encep halitis virus (SLE), and Eastern Equine Encephalitis (EEE). Time placement of the ch icken to their location is critical. A 1991 outbreak of EEE, due possibly to the late plac ement of chickens in mid June and therefore limited vector control, did not receive e arly detection or provide warnings and five human and an above-normal number of horses wer e infected (Day and Stark, 1996).


29 Chickens in a majority of the counties are kept fro m June through December though some year round, and are bled once weekly, or in so me counties every other week, by the mosquito control personnel that keep them (Day and Stark, 1996). After baseline titers are drawn, 1.0ml of blood is drawn and sent to the Department of Health, Bureau of Laboratories in Tampa Florida for serological testi ng for antibodies to the viruses (Day and Stark, 1996). Once a chicken’s sera is determi ned to be antibody positive, the chicken is re-bled for confirmation and if confirme d replaced with a new nave chicken (Day and Stark, 1996). An increase in positive fin dings to 5-10% or more of the flock indicates viral presence and amplification in the v icinity of the flock; it is at this time that mosquito control is implemented to reduce the popul ation of vectors in the area (Calisher et al, 1986). Serological Antibody Detection Method Arbovirus infection can be detected by a variety of serological assays. Serology determines infection past or present based upon the detection of an immune response by the host. Immunoglobulin G (IgG) antibody (Ab) is a monomer with a half life of 23 days, weighing in at 150kDa, it comes in four diffe rent classes 1-4 each varying only slightly in its characteristics; its main functions include opsonization, complement activation, antibody-dependent cell mediated cytoto xicity, and feedback inhibition of B cells (Pier, 2004, Abbas, 2005). In humans and oth er mammals, IgG is an indicator of a past infection, though when that past infection occ urred cannot be determined. IgG has a higher affinity for its protein antigens than immun oglobulin M (IgM). IgM is indicative


30 of a recent infection. It usually appears in humans within 5-10 days of infection and while it has less affinity and specificity for its antige n that IgG, it exhibits greater avidity due to it pentamer or hexamer arrangement of binding sites (Pier, 2004, Abbas, 2005). IgM weighs 950 or 1,140 kDa depending on its arrangemen t and has a half-life of five days at 1.5mg/ml of blood in humans (Pier, 2004, Abbas, 200 5). At the initial stimulation of B cells by infection and the subsequent activation of T-Cells which release important cytokines, IgM begins to be secreted and steadily i ncreases, if this activation continues B cells begin what is called isotype class switching, thereby eventually secreting IgG. These cells are then kept as memory B cells in the event of a subsequent invasion by the same organism (Pier, 2004, Abbas, 2005). It is this pro cess of reactions that serological testing allows us to investigate, in order to determine cou rse of treatment for a patient or activation of epidemiological investigation during surveillance. The following are not exhaustive of the serological methods available for determination of infection. These are however the most commonly applied techniques in the Florida sentinel chicken surveillance program. Hemagglutination Inhibition Test The Hemagglutination test is an economical screeni ng tool for the arboviruses in the Sentinel Chicken Program in Florida. Inhibitio n of agglutination by present antibodies indicates a positive reaction. Positive reactions are seen regardless of whether the immunoglobulin is IgG or IgM. IgM can be detect ed as early as four days post infection, while IgG by day seven (Calisher et al., 1986). IgM titers persist for 250 days


31 post infection, while its titer does drop initially it can then increase again over an eight month period, IgG seen one week after inoculation, peaks at three to four weeks and then subsequently declines (Calisher et al., 1986). Cro ss reactivity is commonly seen in this test between members of the same family; for exampl e, West Nile Virus and St. Louis encephalitis, both from the Flavivirus family and H ighlands J and EEE virus of the Alphaviruses. At the Florida Department of Health the following steps are performed for the HAI testing. This process provides screening result s within one week’s time (samples on Monday or Tuesday and submitted by Wednesday have r esults by Friday) though this does not include confirmation by MAC-ELISA or SN. Processing of sera of the Hemagglutination Inhibiti on testing at the Florida Department of Health follows briefly. o Centrifuge vial at 800 x g for 10min, o transfer supernatant into clean tubes, o aliquot 100uL of serum into new tube set, o place in ice bath and add 0.5mL of Protamine Sulfat e (PSO4), (Holden, 1966) o add 6mL of Acetone and agitate for 5 minutes with w ooden sticks, o pour off acetone and vortex to break up particulat es at the bottom, o repeat last two steps then dry samples overnight. o the next day add 1mL BABS, using sticks to scrape s ides if necessary, o let stand for 1 hour and then centrifuge for 10minu tes at 800 x g, o transfer samples to new vial and add 2 drops of was hed goose erythrocytes, o shake and hold for 20 minutes in ice bath,


32 o centrifuge at 800 x g for 10minutes, o aliquot serum into microtiter well plates, add anti gen, incubate overnight at 4C o add goose erythrocytes, incubate at room temperatur e one hour and read for reactivity. HAI while known to have few false positives it doe s however take time and dedicated personnel. This method has few places in which automation is beneficial and also requires a source of geese erythrocytes. IgM Antibody Capture Enzyme-Linked Immunosorbent As say IgM Antibody Capture Enzyme-Linked Immunosorbent A ssay or MAC-ELISA is a more rapid testing method used for confirmation o f antibodies to Arboviruses following HAI. Time requirement is only two days as opposed t o five with approximately four hours of hands of time for a 40-sample test (Johnso n, 2005). IgM testing is chosen for arboviruses not only for the ability to capture ant ibody in a faster time period, but also IgM is less cross-reactive than IgG (Martin, 2000). The Florida Department of Health Bureau of Laboratories in Tampa briefly uses the fo llowing steps in their Sentinel Chicken sera testing (Martin, 2000). o Add 75uL of anti-chicken IgM capture antibody to mi crotiter plate, o incubate overnight at 4C, o decant antibody and blot, o add 200L of blocking buffer and incubate for 30 mi nutes at room temperature,


33 o wash plate with washing buffer 5 times, o add 50L of chicken sera at 1:400 dilution, add 50 L of positive and negative control sera at 1:400 dilution to the plate, o cover, incubate for 1hr at 37C, o wash plate 5 times, o add 50L of viral antigen to appropriate wells, o add 50L of normal (negative control) antigen to ap propriate wells, o incubate overnight at 4C,. o The following day wash plate 5 times, o add 50L of diluted anti-viral monoclonal antibody conjugated to (name of enzyme) to the wells, o incubate 1 hr at 37C, wash plate 5 times twice, o add 75L TMB (3,3’,5,5’-tetramethylbenzidine) to ea ch well and cover to protect from light, o incubate 10 minutes at room temperature, o add 50L of 1N H2SO4 to stop reaction, o let stand for 1 min and the optical density (OD) is read at 450nm with a spectrophotometer plate reader A serum is determined positive if the P/N is great er than to 2.2 and negative if less than 21.6. A result of equivocal is given if t he result is in the range of P/N 1.6-2.2. Taking the OD of the test serum and antigen and div iding it by the OD of the negative control serum and antigen determines the P/N. A P/N of 2 indicates that the OD of the test sera is twice that of the negative sera. This testing method is one of the methods on


34 which the Luminex Microsphere assay is based, howev er instead of a solid surface (microtiter plate) the surface of reaction is the b ead itself.


35 Plaque Reduction Neutralization Test The Plaque Reduction Neutralization Test or PRNT i s the most specific of the tests thus far discussed for the arboviruses. This test is used in confirmation of positive sera as well as if sera tests negative in the ELISA following a positive reaction in HAI. Though cross-reactivity still occurs it is thought to be less of a problem in the PRNT though ambiguous results still do occur (, 2 004). This test relies upon the actual virus itself for quantification of antibody; it mus t therefore be done in proper BSL-3 containment. This makes the test more technically d ifficult and restricts its use to those laboratories with the proper room, equipment and tr aining. Test serum is titrated and mixed with a known conc entration of test virus. This serum virus mixture (SV) is incubated overnight at 4C and then Vero (African Green monkey kidney) cells are infected with the SV.. Pla ques, areas of cell death occur is a mathematically predictable fashion directly related to the amount of virus plated. Seeing a decrease in the expected number of plaques versus controls indicates that antibody capable of neutralizing the virus exists in the sam ple. Overall it may be two to three weeks before a fina l result is issued on a newly positive or borderline sample. This time delay may cause effects in mosquito control as well as the potential for further animal and human infection. It is a goal of this research to demonstrate that a Microsphere immunoassay for the presence of IgM antibodies to EEE virus will be a sensitive and specific and far more time efficient manner to continue testing for this important virus.


36 Microsphere-based Immunoassay While the MAC-ELISA assays provide faster time to r esult than HAI, a still faster option is now available using new technology. The M icrosphere based immunoassay combines flowcytometry and ELISA technology into one. Instead of a microtiter plate as the solid phase of the assay Microspheres with t wo varying fluorescent dyes, which give up to 100 options of color identification to c hoose from, are used upon which can be bound antigen or antibodies. These beads, through v arying protocols, can be used to quantify such things as antibodies, enzyme substrat es, receptors, viral antigens, and cytokines by using bound fluorochromes to indicate bound anylate. Two of the main advantages of this relatively new technology are sp eed and the ability to multiplex. Assay of a 96-well plate takes between 30-40 minutes and it can be tested for up to 100 different analytes, limited only by the differing beads requi red to indicate the different substances. This large array of different beads allows the test ing of different analytes all in the same well at the same time. In general a Microsphere as say must include a defined bead set, the anylate and a reporter fluorochrome such as R-p hycoerythrin. (Bio-Plex, 2005) The xMAP 5.5-micron polystyrene beads produced by Luminex Corp (Austin, TX) with a carboxylated surface can be bound to suc h things as proteins, oligonucleotides, polysaccharides, lipids, antibodi es, small peptides, antigen or antibodies directly, permitting the testing of the before ment ioned analytes (Kellar and Iannone, 2002). Using varying protocols, samples and contro ls are added followed by the reporter fluorochrome in order to determine the strength of signal in each well for each bead set.


37 The Luminex Bio-Plex system by Bio-Rad draws a samp le using a syringe mechanism from each well and in singular file two lasers are beamed at the sample to excite the fluorescent pigment in the bead for classification and the fluorochrome reporter which determined the fluorescent intensity and further ca n be used in conjunction with a standard to quantify the samples. The first laser is a red laser; also know as a cla ssification laser. This excited the bead and allows determination of the two red fluore scent dyes within and therefore its specificity, which is determined by the researcher. The second laser is a green laser, which is also called the reporter laser. This is us ed to excite the reporter molecule and from this the mean fluorescent intensity of the sam ple is calculated. The MIA protocol currently used by the Florida Dep artment of Health for the Sentinel Chicken sera was designed by Logan Haller M.S.P.H (Haller, 2006) is based upon a duplex method designed by Johnson et al. (20 05) at the DVBID laboratories of the Centers for Disease Control and Prevention for the detection of anti-West Nile and antiSt. Louis antibodies in human sera. Prior to any te sting the carboxylated microspheres (Radix Biosolutions ) must be bound to 6B6C-1 monoc lonal antibodies; this process is completed to two different bead sets which will per mit the differentiation of the two viral antibodies once antigen is subsequently bound the f inal product Positive and negative viral antigens are required for the binding of spec ific IgM antibodies. For example, the protocol for West Nile testing requires the mixing of 580L of WNV positive recombinant COS-1 Antigen with 350L of the bead se t 32 solution and 2870L of running buffer or 160L of WNV negative recombinant antigen, 200L of the bead set 32 solution and 1640 of the running buffer. SLE tes ting requires 35L of SLEV positive


38 SMB antigen, 350L of the bead set 57 and 3115 of t he running buffer, 20L of the SLEV negative Normal mouse brain antigen is added t o 200L of bead set 32 and 1780 of the running buffer. After binding antigen the beads can be used in cap turing specific antibody in the chicken sera. Testing is done in a 96 well filter p late (Millipore cat # MABVN1250, lot# F5HN65106), which allows the reaction to occur and washing steps to be sequentially done and the supernatant vacuumed through the porou s filter bottom (VWR cat# 16003836) without the loss of the 5.5uM beads. The curre nt WN/SLE protocol includes many wash and reaction steps.


39 Objectives As the climate and landscape of our world changes, and people encroach further upon sylvatic habitats, vector borne diseases are b ecoming an ever-greater public health concern. Eastern Equine Encephalitis virus has a mo rtality rate of 30% of those cases diagnosed, while 30% of those surviving infection r emain with neurological sequelae for life (, 2007). The use of sentinel chickens for surveillance of a rboviruses that are known to use birds as a reservoir host, such as St. Louis Enceph alitis (SLE), West Nile (WN) virus, Eastern Equine Encephalitis (EEE) and Highlands J ( HJ) virus, in Florida began with the Sentinel Chicken Arboviral Surveillance Network in 1978 (Day and Stark, 1996). This network enables the activation of an early warning system for citizens, as well as, county epidemiologists and those in mosquito control, allo wing for a coordinated effort of disease prevention. The current methods in use at the Florida Departme nt of Health, Tampa Branch Laboratory include screening of sera for antibodies to these arboviruses of epidemiologic importance by way of the hemagglutination inhibitio n test (HAI), and confirmation by the IgM antibody capture enzyme-linked immunosorben t assay (MAC-ELISA) and Plaque Reduction Neutralization test if the MAC-ELI SA proves to be negative. While


40 these tests combined are providing the results need ed, the time to result can be a week or greater depending on the initial screening result i n the HAI tests. The Microsphere Immunoassay (MIA) provides the cap acity for an improved time to result. In addition, it is expected that the EEE MIA assay will provide results both more specific and sensitive than the current protoc ols. This research aims to demonstrate that the use of a Microsphere-based Immunoassay (MIA) for the detection of EEE IgM anti bodies in chicken sera will provide accurate classification of the results, com parable to those of the ELISA assay, in a significantly reduced time frame from days to hou rs. The specific aims of this research include: 1) determine if Luminex Microsphere Assay Technology c an provide an accurate, more rapid detection method specifically designed to test for antibody to EEE in sentinel chicken sera, 2) determine if any tested samples collected from thos e sera routinely submitted and tested by HAI and shown negative in standard te sting, indicate positive or equivocal result by the MIA method and therefore indicate a missed positive 3) determine the Sensitivity and Specificity, Positive and Negative predictive values of this new protocol as compared with MAC-EL ISA as a reference standard. 4) determine if Highlands J virus will have an impact on the testing protocol and results of this test meant to be specific for E EE


41 Materials and Methods Processing and submission of Sentinel Chicken Sera Samples Serum samples are submitted from counties with chi ckens at sites maintained throughout Florida to the Florida Department of Hea lth Bureau of Laboratories Tampa location weekly. Upon arrival at the laboratory, sa mples are aliquotted for screening using the hemagglutination inhibition antibody test (HAI) for flavivirus (WN and SLE) and alphavirus (EEE and HJ) antibodies (Figure 2). Those screening positive for either of these groups are then submitted for confirmation of IgM using the IgM antibody capture enzyme-linked immunosorbent assay (MAC-ELISA). Foll owing a negative result in the MAC-ELISA the samples are tested using the plaque r eduction neutralization assay, which classifies the sample based on IgG. Following this process the samples are kept at 4C until tested using the Microsphere immunoassay (MIA). In 2007, 2,752 sentinel chickens were maintained in 282 locations throughou t the state. Overall 47,803 sera were submitted for testing to the Florida Department of Health including 5,432 from the southern region, 24,383 from the northern and panha ndle regions and 17,988 from the central region. From these samples 2,162 total ser a were randomly selected for MIA analysis, 748 for the south region, 516 for the nor thern/panhandle region and 681 for the central region. For analysis of known positive HJ p ositive and EEE positive sera, an additional 314 samples were randomly chosen from se ra collected during 2003-2007


42 (Table 2). In addition three wild bird sera were te sted to confirm that this methods primary antibody cannot be used for bird species ot her than chicken.


43 Table 2. Total sera samples submitted during the 20 07-year and samples tested in this research both regionalized by location. r"##$ r% 0123415 & 64755856 'r 31252068 () 9261 145:23376


44 Figure 2. Flowchart for sentinel chicken sera testi ng for antibody to alphaviruses at the Florida Department of Health Bureau of Laboratories Sera are aliquotted separately from the original sample for each test; all samples are stored at 4C. It is first tested for antibody to alphaviruses using the HAI method, whic h has the ability to detect both IgM and IgG. In order to confirm that this bird is newl y infected a MAC-ELISA is performed to determine if IgM is present. If IgM is not found the samples are tested on a PRNT assay to detect any viral specific IgG in the sampl e. Following these standard testing methods, the samples were then assayed using the MI A that detects only IgM. Chicken Sera Tested in HAI HAI positive MAC-ELISA Negative MAC-ELISA Positive MAC-ELISA Confirmation MIA HAI Negative PRNT


45 Sample Size and Specimen selection Due to the regionalized patterns observed in EEE s entinel chicken data (Table 3), samples were drawn based on the region in which the site was maintained (Table 4). On a weekly basis sample numbers were randomly selected using Research Randomizer (, 2008). Sera h ad been stored at 4C prior to testing. To determine the sample size needed for e ach region, the sample size calculator from Cameron and Baldock (1998) was used selecting 80% expected sensitivity and specificity, a level of significance or a = 0.05, a power of 95% and the varying populations and percent of expected prevalence (Tab le 5). Serum analysis Sera for this study were taken from samples submitt ed for arboviral testing at the Florida Department of Health Bureau of Laboratories Positive samples were primarily taken from year 2005 stored samples, though some we re taken from 2003; these positives were chosen with regard to their MAC-ELISA P/N resu lt to incorporate a wide range of values from equivocal (1.6 2.2) to low through hi gh positive (18.6). HAI testing had been performed on these sera using the method by Clarke and Casal (1958), after protamine sulfate and acetone e xtraction were performed to eliminate non-specific protein interactions (Holden, 1966). S era both negative and positive for EEE had been held for testing following the HAI and MAC -ELISA (Martin, 2000) and or PRNT testing at 4C.


46 Table 3. Sentinel chicken seroconversion rates by r egion and state for 2002-2007. Stark, L.M. and Kazanis, D. (2002-2007) Arbovirus S urveillance: Annual Summary Report. Florida Department of Health, Bureau of Lab oratories, Tampa. Year South Central North/Panhandle 2007 0 1.2 7.6 2006 0 0.1 5.0 2005 0.2 2.7 14.5 2004 0.2 1.0 7.8 2003 0 2.1 13.6 2002 0 1.1 4.9 2007 MASR 95% CI* 0.03-0.43 0.46-3.48 7.92-12.16 Confidence interval of the Mean Annual Sero-conve rsion Rate


47 Table 4. Counties by region used for sample selecti on Panhandle North Central South Alachua Citrus Brevard Charlotte Bay Duval Hillsborough Collier Escambia Flagler Indian River Dade Gulf Nassau Manatee Glades Jackson Orange Osceola Hendry Leon Pasco Pinellas Lee Walton* Putnam Sarasota Martin Seminole St. Lucie Palm Beach St. Johns Volusia wild bird sera also submitted but not tested by M IA


48 Table 5. The sample size for each region as determi ned using the sample size calculator from Cameron and Baldock (1998). Sero-conversion me ans were taken from 2002-2006 arbovirus surveillance reports. Stark, L.M. and Kazanis, D. (2002-2006) Arbovirus S urveillance: Annual Summary Report. Florida Department of Health, Bureau of Laboratorie s, Tampa. Region: Mean sero-conversion Mean population Number of samples recommended South 0.03-0.43% 758 758 Central 0.46-3.48% 1008 673 North 7.92-12.16% 1265 479 Panhandle 7.92-12.16% 506 52 1962


49 The process for MAC-ELISA as previously discussed using antigen provided by CDC, was performed by technologists of the Florida Department of Health, Tampa Branch. Briefly lyophilized goat anti-chicken antib ody (MP Biomedicals, cat# 64395, lot# 8155H) diluted at 1:1000 in buffer (0.015M sod ium carbonate, 0.035 M sodium bicarbonate, pH 9.6) and placed in separate 96 well microtiter plates. Sera and controls diluted 1:400 in wash buffer were then added in dup licate and EEE Antigen (CDC, cat# M29603) was added at 1:800 dilutions in wash buffer After an overnight incubation at 4C, monoclonal antibody to EEE virus conjugated to horseradish peroxidase (CDC, cat# VS2371) was added. After a one-hour incubation at 37C, the plate was washed and the enzyme substrate TMB (3,3’,5,5’-tetramethylbenzidin e) (Sigma, cat# T8665), was added. After 10 minutes, the reaction was stopped with 1N H2SO4 and the plates read for their optical density (OD) at 450 nm, using a Beckman Cou lter AD340 spectrophotometer. Positive to negative ratios (P/N) was determined by dividing the mean OD of the test serum with the viral antigen by the mean OD of the negative control serum on the same plate. If the P/N is greater than or equal to 2.2 t he serum is considered positive, if 1.6 to 2.2 it is considered equivocal and less than 1.6 it is considered negative. Equivocal and negative results are then confirmed using the PRNT method. Samples that are negative for antibodies by HAI are not tested further in the MAC-ELISA therefore they were tested directly using the Luminex. For those samples chosen to be included in the calculation of the cut off values, MAC-ELISAs were performed to determine IgM value. A sample of 720 s era testing both positive or negative in MIA as well as MAC-ELISA were then used in the d etermination of result cutoff using


50 Receiver Operator Characteristic (ROC) curves from Analyse-it software v2.0 (AnalyseIt Software LTD. England, United Kingdom).


51 Microsphere-based Immunoassay Addition of Antigen to Bead Sets Three m l of normal suckling mouse brain (NSMB, CDC cat# M2 9714) or EEE+ mouse brain antigen (CDC, cat# M29603) and running buffer, phosphate buffered saline, with BSA, pH 7.4, was combined with bead set 15 cou pled to 2A2C-3 monoclonal antibody to alphavirus (Radix Biosolutions, Georget own TX) (Johnson (CDC, Personal Communication) (Table 6). The mixture was placed in a 4ml brown Nalgene bottl e (Nalgene, HDPE lot #2004-915) to limit light absorption and incubated on a labquake tube rotator at 8 RPM (VWR cat # 56264-302) for one hour at room temperat ure. This method gives a final concentration of 500 beads/L. After antigen captur e, the beads were kept at 4C for no more than one month (Johnson, 2005). Working diluti ons of bead sets, primary and secondary antibodies were prepared on the day of ex perimentation, kept on ice and wrapped in foil to protect from ambient light bleac hing of the bead sets and the reporter molecule prior to use.


52 Table 6. Volume of reagents used to create antigen coupled bead set 15 at 500 beads/L with a total volume of 1000 m l. Reagent Volume in m mm m L Buffer 897 NSMB or EEE+ Antigen 3 Bead set 15 with 2A2C-3 antibody 100 Total volume 1000


53 Standardization of Microsphere-Based Immunoassay Testing of human sera for IgM requires the depleti on of IgG from the sample. To rule out the need for such treatment for chicken se ra, multiple samples were subjected to IgG depletions using the Protein G sepharose 4 Fast flow (Amersham Biosciences #NC9354476) following the manufacturer’s recommenda tions. Briefly, a slurry was made using one part Protein G sepharose (PGS) to th ree parts 20% ethanol. A filter plate was pre-wet with PBS for five minutes. After suctio ning off the PBS, 80 m l of PBS plus 20 m l of PGS slurry were added and then the diluent imm ediately suctioned off (Multiscreen Resist vacuum manifold cat#MAVM0960R); 95 m l of PBS and 5 m l of neat serum were added. Mixture was resuspended by shakin g with a Lab-line plate shaker platform (VWR# 57019-600) at 1100 RPM. A microtiter plate was placed below the filter plate to collect the now IgG depleted serum from th e filter plate after vacuum filtration. In order to determine if this process was necessary in chicken sera MIA assay results of depleted and non-depleted sera were compared using an ANOVA. Testing chicken sera as opposed to human sera offer s some challenges. For example, in the research by Johnson et. al with hum an clinical samples, there was available a single anti-human IgM antibody conjugat ed to the detector molecule phycoerythrin (PE) commercially available, however, such a commercial product is not available for chicken antibody and therefore a two antibody approach with goat antichicken and porcine anti-goat conjugated to PE was used (Haller, 2004). This deviation from the original method designed by Johnson et al. provided an economic and feasible alternative.


54 The specially designed filter plate (Millipore cat # MABVN1250) was divided in two halves, the left containing the positive antige n bead set and the right containing the negative antigen bead set. In order to determine th e best concentration for the bead set, using the least number of beads, three concentratio ns of bead set with buffer were tested and the data analyzed using student’s t-test and AN OVA. Bead concentrations included a 1:5, 1:10 and 1:20 dilution of the stock solution t ested at various serum concentrations (1:80, 1:160, 1:320, 1:400, 1:640, 1:1280) in dupli cate, holding the primary and secondary antibody dilutions at 2g/ml and 1g/mL r espectively, a dilution previously determined optimal for the WN/SLE test by Haller. T hese findings were then tested using an ANOVA followed by Tukey’s multiple comparison te sts. Determination of the most appropriate dilution for the chicken sera using the least sample was accomplished in a similar manner, 24 kno wn positive and negative sera were tested in duplicate and then compared at 1:80, 1:16 0, 1:320, 1:400, 1:640, 1:1280 between each concentration by ANOVA followed by a T ukey’s multiple comparison test. These dilutions were performed holding the primary and secondary antibody dilutions at 2g/ml and 1g/mL respectively. The optimal concentration of the primary goat anti -chicken IgM serum lyophilized (MP Biomedicals, cat# 64395) antibody a nd secondary porcine anti-goat IgGphycoerythrin (PE) (R&D Systems, cat# F0106) antibo dy was done by testing 23 samples in duplicate using three different combinations, 4 g/mL of primary goat anti-chicken with 2g/mL of porcine anti-goat PE, 2g/mL of prim ary goat anti-chicken with 2g/mL of porcine anti-goat PE and finally 2g/mL of prima ry goat anti-chicken with 1g/mL of porcine anti-goat PE all at 1:400 sera dilution fac tor; These were then evaluated using an


55 ANOVA followed by a column chart of the values and a Tukey’s multiple comparison test. In order to determine that antibody binding was giv ing true reading results, not aberrant noise and that incomplete test wells would not give false results, tests were run for a baseline of fluorescent response with positiv e and negative known samples with and without primary as well as secondary antibody (Tabl e 7). For the establishment of the cutoff values using RO C curves, 720 specimens were selected with both Luminex and MAC-ELISA results. T he method of data transformation (P/N) method used in the ELISA was chosen to determ ine the cutoff values at the screening level. To determine the cutoffs, the posi tive Ag beads set Mean Fluorescent Intensity (MFI) value was divided by the Negative c ontrol sera MFI value of the plate, this was considered the transformed value for scree ning. The same cutoff value was applied to the confirmation testing which incorpora ted the division of the positive antigen bead set MFI divided by the negative antigen bead s et MFI for that serum. This confirmation, which took into account the backgroun d reactivity of each sample, was critical to determine if the sera were indeed react ive to the antigen and not to something non-specific in the well. Results discussed later w ill show the importance of the negative antigen value when determining the final result con clusion for each serum.


56 Table 7. Combinations employed to confirm absence o f reactivity in incomplete wells. Sample Primary Secondary None Present Present Present None Present Present Present None Present Present Present


57 Microsphere-Based Immunoassay Protocol Bead sets conjugated with positive viral or negati ve antigen as previously described were diluted 1:10 with running buffer; se ra were diluted at a 1:400 in deep well plates, with running buffer. Primary and secondary antibodies were diluted with running buffer at 2g/mL and 1g/mL respectively; all worki ng diluent dilutions were covered with aluminum foil and kept on ice for use the same day. To establish the most accurate assay with the least cost both a screening and conf irmation protocol were designed. In the screening protocol a 96 well filter plate w as designed with a negative and positive control in duplicate, these controls are a lso found on the confirmation plate on both the positive antigen and negative (control) an tigen sides. The goal of screening was to establish a preliminary reference for the number of sera with potential to have confirmed results. In this case the MFI values of e ach tested sera were divided by the negative control MFI values. The negative control s erum used had been through the standard process on negative confirmation as well a s the MAC-ELISA prior to use in this and other assays requiring a sera negative control. The 96 well plate was divided into two halves for the confirmation protocol; one for the reaction with the positive EEE antigen (CDC cat# M29603) (left) and one for the negative NSMB antigen (CDC, cat# 0006) (right) (Fig ure 3). This configuration permitted the assay of 44 sera samples per plate. T o hydrate the filter, 100L of running buffer was added to each well and left in place for five minutes; 50L of each bead set (diluted 1:10) was added to its respective side and the plate washed twice with running


58 buffer. 50L of sera (dilution 1:400) was then adde d and the plate, covered with tin foil to protect from ambient light bleaching, was placed on a Lab-line plate shaker platform (VWR# 57019-600) initially for 30 seconds at 1100rp m and then for one hour at 300 rpm. The plate was washed twice with running buffer primary antibody 2g/mL was added and the plate was incubated on the plate shak er for 30 seconds at 1100 rpm and then at 300 rpm for thirty minutes. After the plate was again washed twice with running buffer, secondary antibody was added at 1g/ml and the plate was incubated at room temperature for 30 minutes on the shaker at 300 rpm after 30 seconds at 1100 rpm. After a final double wash, 100l of running buffer was ad ded and the plate shaken at 1100 rpm for 30 seconds. This completes plate preparation; i t is placed into the Luminex Bioplex 100 suspension array system (Bio-Rad Labs (BioPlex) cat# 171-203060) for assay. The software returns the results in MFI (Mean Fluoresce nt Intensity). During setup of the instrument protocol, settings different than factory defaults were used, these were: 100 beads per region were re ad with a sample size per needle draw of 75 m l. The Override gates were placed at 5,000-10,500. Results were then selected to be automatically stored; worksheets wer e categorized by bead set. Discernment of IgM Antibodies to EEE Prior to each assay run, the Luminex system was tu rned on and the lasers warmed up for one half hour and the level of Bioplex sheat h fluid (BioPLex, 20L cat#171000055) was checked to ensure testing could be completed wi thout interruption or artifact due to


59 clumping or bubbles. After warm up, instrument cali bration was performed and data relating to the samples and dilution were entered i nto the protocol screen. Each bead set in a Luminex microsphere assay is re ad first by the red laser, also known as the classification laser. This laser quant ifies the number and type of bead set in the sample indicating the number and percent of clu mping of sample in the results. The green laser, also known as the reporter laser, quan tifies the R-Phycoerythrin attached to the secondary antibody, porcine anti-goat. The inte nsity of the R-PE signal is directly proportional to the concentration of the PE that is bound to the chicken antibody/antigen /2A2C-3 antibody/bead complex. This value is then t ranslated by the software into the Mean Fluorescent Intensity (MFI) of the samples


60 EEE positive Antigen Bead Set N SMB Control (Negative) Antigen Bead Set EEE positive sentinel chicken control sera EEE negative control sera Sentinel Chicken test sera Figure 3. Confirmation Plate design used for testin g of chicken sera for IgM antibodies. Original design for plate illustration taken from w ork by Haller (2006).


61 Classification of Luminex MFI Results Classification of the results was based upon a tra nsformation of the Mean Fluorescence Intensity (MFI) data. The transformed values for screened sera were calculated by using the standard method used also i n the MAC-ELISA testing algorithm known as a P/N value. In brief the MFI of the negat ive control, tested on the positive bead set, was divided into the MFI of the sera test ed on the positive antigen alone. The transformed value in the confirmation protocol vari ed slightly due to the addition of the testing of negative Ag bead sets. This negative ant igen bead set provides an assessment of the reactivity of the individual sera against th e proteins of the NSMB; it allows the calculated removal of non-specific reaction by the antibodies, this therefore takes into account any reaction other than the reaction to the viral antigen. By dividing the positive antigen bead MFI by the reaction to the negative an tigen bead set the non-specific reactivity is removed and its value can be compared to the cutoff and assessed for final interpreted result. Receiver Operating Curves (ROC) were employed to de termine an appropriate cutoff value using this method. ROC curves were gen erated with the Analyse-It Software, a Microsoft add-in program (http://www.analyse-it.c om). These tests allow a visual interpretation of the assay’s ability to discern be tween positive (antibody detected) and negative (antibody not detected) populations, in th is case chickens. By analyzing the area under the curve of the two normal curves produced b y these different populations and following the corresponding listing of the true pos itive and true negative fractions


62 (sensitivity and specificity) the best cutoff value of 9.7 was determined and subsequently used for result determination. In this instance 720 confirmation transformed valu es of positive and negative sera were compared against the final positive or negativ e results from the MAC-ELISA, the closest to a gold standard for IgM available at thi s time. These results were evaluated and the cutoff and test sensitivity and specificity, po sitive predictive and negative predictive power noted. A total of 1,590 further samples were then evaluat ed using these cutoffs; 38 of these samples were known Highlands J (HJ) antibody positive. Testing of HJ positives were done to show that this other Alphavirus common ly cross-reactive in the HAI would not affect this test. In addition three wild bird s amples were also assayed to re-affirm the specificity of the test to chicken antibodies.


63 Results Due to the regionalization of EEE virus in Florida sample size was determined using the mean sero-conversion rates and the mean p opulations of each testing region. Sample sizes were determined by the method of Camer on and Baldock to be 758 for the southern region, 673 for the central region, and 47 9 for the combined north and panhandle regions. The total number of samples test ed from those found negative by HAI for each region included 681 for the central region 748 for the south region, 516 for the northern/panhandle region, and 32 from samples take n in Alabama, in addition to this 314 samples known to be positive for WN/SLE, EEE, a nd HJ by HAI from the years 2003, 2005, 2006 and 2007 and 3 from wild birds. Luminex Microsphere Immunoassay Technology Protocol design for this Luminex assay required th e evaluation of dilution factors for the sera, bead sets, primary and secondary anti bodies. Prior to determining these values however it was first necessary to evaluate t he need for IgG depletion of the chicken sera as this is a common practice in human sera. Cutoffs were then determined and all results were analyzed, those with Luminex r esults differing from the HAI and MAC-ELISA results were scrutinized. The sensitivity specificity, positive and negative predictive values were also evaluated compared to b oth the HAI as well as the MAC-


64 ELISA results. Lastly due to the cross reactivity o f the alphaviruses, sera known to be positive for antibody to Highlands J virus, sometim es found in the Florida sentinel chicken sera by the DOH, was also tested to indicat e whether the antibodies to this virus might give a false positive for EEE in the MIA. IgG Depletion of Sentinel Chicken Sera Previous Luminex microsphere assays in humans for IgM have depleted the sera of IgG prior to testing (Johnson et al, 2005), whil e chickens technically have an immunoglobulin known as IgY with slightly different characteristics that IgG it was still important to determine if this immunoglobulin cause d any interference in result. Therefore to determine if such IgG depletion steps are necessary in the testing of chicken sera, 40 sera were testing with and without depleti on. The data was then transformed into a P/N ratio by taking the positive antigen test ser um MFI and dividing by the positive antigen MFI of the negative control serum. This tra nsformed data was then analyzed using an ANOVA table (Table 8). The F value was fou nd to be significantly less than critical F and the P-value also indicated that the differences between the depleted and non-depleted sera were not significant. All samples tested following this determination were no longer treated to remove IgG.


65 Table 8. Single Factor ANOVA table for depleted vs. non-depleted chicken sera Anova: Single Factor, transformed data for depletio n vs. non-depletion of sera Groups Count Sum Average Variance transformed data depletion 40 1383.423 34.585 1413.998 transformed data no depletion 40 1331.205 33.280 1325.082 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 34.084 1 34.084 0.0248 0.875 3.963 Within Groups 106824.1 78 1369.540 Total 106858.2 79


66 Dilution Factors for Testing Protocol Bead set dilutions were evaluated based on 24 sampl es at 1:5, 1:10 and 1:20 dilutions. An ANOVA of the data between dilutions i ndicates that since the F 1.35 is less than the F critical 3.12 and the p-value is 0.26 th at the dilution of the bead set has no significance at the varying sera dilutions of 1:80, 1:160, 1:320, 1:400, 1:640 and 1:1280 (Table 9). In order to confirm the ANOVA finding a Tukey’s multiple comparison test was done. The Tukey’s multiple comparison test also known as the Tukey’s honestly significant difference test or HSD is one available test for use in determining which means amongst a set of means that differ from the r est being compared. With only two groups being compared the t-test would be sufficien t, however, when comparing more than one mean this method would be inappropriate. T he Tukey HSD test like the t-test and ANOVA assumes that data from the different grou ps are from different populations, have normal distribution and the same standard devi ation in each group (Table 10).


67 Table 9. Single Factor ANOVA table for bead set dil utions at sera dilutions of 1:80, 1:160, 1:320, 1:400, 1:640 and 1:1280. The F value of 1.35 < critical F 3.12 and therefore indicates that there is no difference between the b ead set dilutions. The P-value also reflects this result.. Groups Count Sum Average Variance Bead set 1:5 24 17104 712.6667 1116462 Bead set 1:10 24 16355.5 681.4792 977367.9 Bead set 1:20 24 7945 331.0417 299771.8 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 2155334 2 1077667 1.350685 0.265825 3.129642 Within Groups 55052831 69 797867.1 Total 57208164 71


68 Table 10. Tukey’s multiple comparison test for the bead set dilutions 1:5, 1:10 and 1:20, calculated from P/N transformed data with positive and negative antigen. These results indicate that there is no significant difference am ong the transformed values between the three different beads set dilutions. Therefore a 1: 10 dilution was chosen to coincide with the dilution used in the WN/SLE testing currently i n place to simplify future multiplexing efforts. Dilution Difference between means 95% Confidence Li mits Positive Antigen bead set 1:5 1:10 66.6 -997.5 1130.6 1:5 1:20 768.8 -295.3 1832.8 1:20 – 1:10 702.2 -361.8 1766.3 Negative (NSMB) bead set 1:5 – 1:10 4.2 -88.1 – 96. 1:5 – 1:20 5.5 -86.7 – 97.8 1:10 – 1:20 1.3 -90.9 – 93.6


69 Sera dilution for the MAC-ELISA and current WN/SLE microsphere assay are currently at 1:400. To evaluate if this dilution wa s also appropriate for the EEE microsphere assay, 24 serum samples, 144 separate o bservations of known positives and negatives, were assayed at varying dilutions. ANOV A resulted in an F value less than critical F 0.14 > 1.82 and a P-value of 0.98, thus it was determined that all dilutions were equivalent. A Tukey’s multiple comparison test was then completed to confirm there was no point of dissimilarity between the dilutions. Tu key’s HSD indicated that all sera dilutions were grouped together in the same Tukey g rouping as well as all 95% confidence intervals included the value of zero. Ba sed on the ANOVA and the Tukey’s multiple comparison procedure results that no speci fic dilution was preferred the decision was to use the standard of 1:400 as is found in the MAC-ELISA. Before establishing this value however all samples for bead set, and primary and secondary antibody dilution were tested with these multiple dilutions and then evaluated based upon the 1:400 after dilution for the sera was determined. Antibody dilution for the primary goat anti-chicke n IgM and the secondary porcine anti-goat IgG-PE when evaluated by ANOVA in dicated that the dilutions were not all equal, with a F value of 67.34 > F critical of 2.06. To indicate which treatment varied from the null hypothesis a Tukey’s multiple comparison test was done. The Tukey test indicates significant differences in a variety of the different dilutions (Table 11a and Table 11b) it was however obvious from the grouping s that the dilutions were divided by the strength of the secondary antibody. By evaluati ng the data individually in a column graph we can easily visualize the differences found in the Tukey examination (Figure 4 and 5). It can be seen that the point of saturation for the primary antibody can be found at


70 1 m g/ml, this is indicated by the lack of change in si gnal strength at this dilution and it is in fact the secondary antibody that changes the out come of the signal strength due to its saturation of the binding points found on the prima ry antibody. Looking specifically at the transformed MFI’s on the Y-axis a value of 40 c an be seen for the 2g/ml of primary with 1g/ml of the secondary. This is an acceptable value and indicating saturation by the primary antibody but less that saturation of the se condary antibody at that dilution. These dilutions permit easy differentiation of positive a nd negative reactions. This dilution was also deemed appropriate in the development of the M IA protocols for detection of antibodies to WN and SLE viruses in chicken sera. T he future goal is to multiplex this assay with the WN and SLE protocols and it was this that therefore became a deciding factor in limiting the dilution of the primary and secondary to 2g/ml of the primary and 1g/ml of the secondary.


71 Table 11a. Tukey Comparisons for significance of di fference between various Ab dilutions using P/N transformed MFI values. Values with significant difference at the 0.05 level are indicated by ***. Dilution Comparison Between Means Simultaneous 95%Confidence Limits 3 – 5 3.796 -12.352 19.944 3 – 1 4.272 -11.876 20.419 3 6 44.982 28.834 61.130 *** 3 2 47.968 31.820 64.116 *** 3 4 50.199 34.051 66.347 *** 5 1 0.476 -15.672 16.623 5 6 41.186 25.038 57.334 *** 5 2 44.172 28.024 60.320 *** 5 4 46.403 30.255 62.551 *** 1 6 40.711 24.563 56.859 *** 1 2 43.697 27.549 59.844 *** 1 4 45.927 29.779 62.075 *** 6 2 2.986 -13.162 19.134 6 4 5.217 -10.931 21.365 2 4 2.231 -13.917 18.379 4 6 -5.217 -21.365 10.931


72 Table 11b. Tukey comparison groupings for significa nce of difference between various Ab dilutions using P/N transformed MFI values. Dilution Factor Tukey Grouping Mean N Dilution Num 2 g 1 / 2 g 2 A 92.520 7 3 1 g 1 / 2 g 2 A 88.724 7 5 4 g 1/ 2 g 2 A 88.249 7 1 1 g 1 / 1 g 2 B 47.538 7 6 4 g 1 / 1 g 2 B 44.552 7 2 2 g 1 / 1 g 2 B 42.321 7 4


73 Figure 4. Column graph of primary and secondary ant ibody dilution with viral antigen positive bead sets. The Y-axis is the P/N transform ed value while the X-axis shows the dilutions of primary and secondary antibodies analy zed. The testing of sera completed after analysis was kept at the value of 2 m g/ml of primary and 1 m g/ml of secondary (2 g 1 / 1 g 2).


74 Figure 5. Column graph of primary and secondary ant ibody dilution with EEE negative control bead sets. The Y-axis is the P/N transforme d value while the X-axis shows the dilutions of primary and secondary antibodies analy zed. The testing of sera completed after analysis was kept at the value of 2 m g/ml of primary and 1 m g/ml of secondary (2 g 1 / 1 g 2).


75 Classification of the Microsphere–Based Immunoassay Result To determine the cutoff values for the Luminex micr osphere assay for EEE, 720 samples, 530 negative and 190 positive, were tested by HAI, MAC-ELISA, and then MIA. The positive or negative results from the MAC-ELISA wer e paired with transformed MFI values of the same sera and analyzed by the Analyse it software (Figure 6). Sera with equivocal IgM ELISA results were treated as negativ e since they have a P/N < 2. MIA data was transformed into a P/N value by dividing t he Positive antigen MFI by the negative control value on the same plate. The resul ting ROC plot illustrates a cutoff value for the transformed data of 9.7 with a sensitivity of 97%, specificity of 95%, a positive predictive value of 87% and negative predictive val ue of 99%. The confirmatory test data was also held at this same cutoff though transforma tion of this data also involves the division of the positive MFI value by the MFI value of the negative bead set for the same serum. Once calculations on all tested samples were complete, samples with disagreement between the microsphere assay and the HAI and MAC-ELISA results were noted and explanation for discrepancies sought.


76 Figure 6. EEE ROC curve showing a visual representa tion of the cutoff value (9.7). The sensitivity can be seen on the y-axis and 1-specifi city along the x-axis along the bottom. The sensitivity of the assay at a cutoff value of 9 .7 is 97% while specificity is 95%, with a positive predictive value of 87% and negative pre dictive value of 99%


77 Detection of Antibodies to Eastern Equine Encephali tis In order to gauge the accuracy of the MIA assay, H AI and MAC-ELISA testing was done in conjunction with MIA and the results co mpared. MIA assays were performed on 2,290 specimens including 720 also ass ayed with the MAC-ELISA. Of these samples, nine were included that were flavivi rus positive to gauge any crossreactivity that might stall future efforts of multi plexing with the WN/SLE test, 40 HJ positive sera were included to define any alphaviru s cross-reactivity and three wild bird sera to show chicken specificity. When evaluated, 9 2 samples were found to have results from the MIA that disagreed with the HAI through th e screening method and 75 that differed in the confirmation results (+Ag vs. –Ag). Results were analyzed to discern a possible explanation for the observed differences. A majority of the samples with HAI to MIA disparity had an ELISA value that supported the MIA result. Samples labeled “unconfirmed” are due to the lack of testing of th e sample using the confirmation protocol. In other words sample was not available t o run on a +Ag/-Ag plate which could then be compared to the P/N screening result. Sera that were equivocal on ELISA testing were also considered and 50% of those tested with t he MIA were found to be classified positive while the other half were classified negat ive by this new method. The MIA assay detected IgM in sera several weeks a fter the initial positive assay, the longest tested at four weeks post first positiv e ELISA result; these counted with the true or late positive group. Additionally, the MIA detected missed samples, that is, samples testing negative in HAI and ELISA but found to have MFI levels well above the cutoff level not due to high background. Comparing the MAC-ELISA results after


78 removal of the 14 MAC-ELISA equivocal, 40 HJ and 3 wild sera, 31 discrepancies were noted out of 706 with the screening method and 41 w ith the confirmation method. In order to evaluate if the MIA assay was incorrect in these instances, MIA results were compared to the true results (arrived at by interpr eting the HAI, MAC-ELISA and PRNT together) and attempts were made to explain them. When comparing the ELISA with the screening P/N va lue, 15 of the 31 samples were found to have true results of EEE positivity, this includes newly positive birds as well as sera that were from bird previously positiv e for antibodies (two to seven weeks). Four of the results could not be confirmed as posit ive due to the lack of specimen available to retest for using the confirmation prot ocol result by +Ag/-Ag transformation, three samples were equivocal on the ELISA and since MIA does not have this classification, positive or negative was assigned, but similar to the HAI two were positive and one was negative. Excitingly possibly one more HAI missed positive was found. The ELISA vs. MIA confirmation testing with +Ag/-A g had 30 samples where the MIA agreed with the HAI result: 17 were found t o be true or late positives which the ELISA did not pick up on, due possibly to the low I gM levels; eight were equivocal on ELISA and were divided with 50% showing positive in the MIA to 50% showing negative results when tested by the MIA. Three of t he samples appeared to have high negative antigen background and were therefore a fa lse MIA positive and two 2005 samples were found negative by MIA, possibly becaus e they were stored only at 4C for an unknown period of time and had lost some IgM in that time. Analysis of the difference between the screening a nd confirmation testing was performed to evaluate reasons for discrepancies. Ni ne screening and confirmation tests


79 did not agree. Five were found to be positive with screening and negative upon confirmation, of these two sera had a true value re sult of EEE positive and two negative. Of those testing negative on the screening four wer e found to be positive after confirmation testing, all of these had a true value of positive, two of which were known to be greater than 2 weeks post-infection. HAI Negative Samples Showing Luminex Positivity Of the 1,976 HAI negative samples tested 28 were f ound to be positive by MIA screening methods. ELISA found eight of the 28 scre ening results negative, but ten of the samples, three of which, were also negative by ELIS A, were found positive by the confirmation method. There was insufficient serum r emaining for 22 of the samples to be tested by the confirmation method and therefore the y could not be confirmed to have a valid positive. Of those that were tested by both t he screening and confirmation methods, five had all testing methods with which to make an assessment of the reasoning behind the result (Table 12). One though screening positiv e, was found negative by confirmation; this may be due to either a strong ba ckground reaction or that the sera was truly negative and was caught with a strong positiv e screening result due to a low negative control sera. Two samples were found to be from a previously positive bird, MIA was more sensitive to the IgM in the sera than the ELISA and therefore was able to detect its existence. There are two positive sera t hat were previously mentioned to be HAI missed positives. The strong positive values (b oth markedly greater than the 9.7 cutoff) lead to the conclusion that since this test may be far more sensitive than HAI or


80 ELISA that these positives may previously have been missed. Both of these missed sera are from counties with multiple positive birds thro ugh the year 2007.


81 Table 12. Sera found negative by conventional metho ds, found positive in the MIA. There are varying re asons for this situation to occur. In two of these cases due to the strength of the signal shown in the MIA it is possible that po sitives have been missed using HAI and MAC-ELISA. In more instances than listed he re sera greater than two weeks post infection (up t o seven) were tested and all were found to be determined positive by either the screening or confirmation methods. %* &+,%!* %%!n %.n-n %&/* % nnn3:'1;r3'0n n nnn60'1;r66'5;r <= nn9 63'0 ;r 36'7 ;r nnn66'2;r6:'6;r ()> nnn 64'6 ;r 2:'0 ;r


82 Assay Sensitivity, Specificity, Positive Predictive Samples were first analyzed for results with HAI, MAC-ELISA and MIA to determine a cutoff value for the transformed data. The best cutoff value of 9.7 was determined by the Analyse it software (2007). Follo wing the establishment of the cutoff value it was incorporated into data analysis to det ermine the result of the screening and confirmation transformed data values and from this the final conclusion was developed. In order to assess the sensitivity, specificity, P PV, and NPV of the MIA screening and confirmation tests against the other testing me thods, they were compared with the results found from the testing of samples by HAI as well as MAC-ELISA (Table 13). Sensitivity of the MIA was 65% to 67% when compared to the HAI test results for the confirmation and screening results respectively. Sp ecificity however was 99% accurate in both the screening and confirmation methods. PPV wa s found at 88% and 95% for the screening followed by the confirmation and lastly t he negative predictive values were 95% and 90% for the same methods. The MIA proved to have a strong screening and conf irmation sensitivity of 97% and 86% when compared to the MAC-ELISA, the current standard for IgM testing in sentinel chicken sera. The specificity remained hig h at 95% and 94%, and negative predictive value at 99% and 95%. Positive predictiv e value for the screening however remained the same as the comparison to HAI at 88% w hile the confirmation decreased to 84%.


83 Table 13 Sensitivity, Specificity, Positive and Neg ative predictive values for EEE MIA protocol vs. the HAI and ELISA methods. These value s are calculated between the HAI and MIA screening and confirmation and the ELISA ve rsus the screening and confirmation. %0%+,%!0% !n &/ !n &/ :'84:'80:'74:'58 !/ :'77:'77:'70:'71 1 :'55:'70:'55:'51 1 :'70:'7::'77:'70


84 Testing Results of the cross-reactive Alphavirus Hi ghlands J Highlands J is an Alphavirus known to be cross-rea ctive with EEE antibodies. Highlands J positive samples from the north and cen tral regions, collected in 2005 were analyzed to determine if the MIA assay could distin guish between antibody to this virus and antibody to EEE. Forty known HJ positive sample s in total were run, all had tested positive for alphavirus in HAI and confirmed negati ve for EEE by ELISA and assigned a true result of positive for HJ through PRNT. Of the se 40 samples, not one was found to be positive by the MIA assay through the screening method. There were, however, three that tested positive in the confirmation assay, whi le two were close to the cutoff at 11.1 and 16.3 one was found to be more than twice the cu toff at 20.5. Due to the need that both the screening and confirmation result combined be positive for a positive final result for the MIA both of the sera mentioned above were d etermined negative. This means that the MIA correctly determined through the combinatio n of screening and confirmation the correct result in all 40 specimens known to be HJ p ositive. It is possible that the one serum that showed a P/N of 9.2 and a positive confi rmation was in fact misidentified in 2005 and that after this 2 year time period the IgM has in fact diminished in the sample.


85 Discussion This study was undertaken to develop and prove the efficacy of a microsphere immunoassay (MIA) for the quantification of antibod ies to Eastern Equine Encephalitis virus in chicken sera. Counties that participate in the Florida Sentinel Chicken program, which has been in operation since 1978, submit sera weekly. Currently sera are tested using the HAI followed by the MAC-ELISA followed by then PRNT. This flow of testing and results provides preliminary HAI result s the same week of submission and confirmatory results in 2-3 weeks. The MIA assay on the other hand takes a total preparation time of approximately three hours and a running time of 45 minutes per plate and requires only a minute amount of sample. This e stimate doe not include the accession of the sera that is currently done for the HAI test ing and would still be done prior to testing with this method. However with the inclusio n of automation working together with the technicians of the laboratory this still h as the potential to decrease time to result greatly, which would then impact the speed at which individual counties could implement actions to address the mosquito vector and warn res idents. Of even greater importance in the implementation of this method is the increase i n the sensitivity of the method which will increase true positives and decrease false pos itives and therefore decrease the amount of time and cost necessary for confirmatory testing


86 Johnson et al (2005) previously designed a study f or the detection of WN and SLE viruses in human sera. This was followed by the work by Logan Haller (2006) for an MIA assay for antibodies to the Flaviviruses WN and SLE for the Florida Sentinel Chicken program sera at the Florida Department of H ealth. Haller was able to adapt the human sera assay reagents and dilutions to establis h a sensitive and specific test for WN and SLE. EEE is a serious disease for humans as wel l as horses and up until this study antibodies to it in chicken sera had not yet been t ested for using the MIA assay. The first aim of this research was to determine if the MIA assay technology could be implemented and shown to be a more rapid and sti ll accurate method of testing for antibodies to EEE. The research began by determinin g if depletion, such as in the work by Johnson et al. (2005) with human sera, was neces sary. This was found to not be the case in both the research by Haller (2006) as well as in this current method. Dilutions of the reagents involved in the MIA meth od were also tested for the best result with least background, using the least quant ity of reagent. Initially three microliters of antigen or NSMB were mixed with the bead set 15 with 2A2C-3 coupled antibodies (Radix Biosolutions) and 100l of running buffer (Johnson, personal communicatio n). Sera were found to produce a sufficiently strong si gnal-to-noise result when assayed at a 1:400 dilution in running buffer. Thus four microli ters of sera were combined with 1,600 l of running buffer. The optimal dilution of Lumin ex bead set 15 with positive and negative antigen was a 1:10 dilution with running b uffer prepared on the day of test run, to a volume sufficient for 50l per well. Haller (2006) used a two-antibody combi nation in order to bind the reporter molecule (R-PE) to th e captured chicken antibody which provides the quantification of bound chicken IgM. T his is in contrast to the work by


87 Johnson et. al (2005) who was able to use a single antibody, directly linked to the R-PE for assay of human sera. The same two-antibody appr oach was also used in this assays development. The primary goat anti-chicken IgM and the secondary porcine anti-goat IgG-PE were found to be at their best at 1g/1ml, t he primaries point of saturation, and 2g/ml the dilution with the strongest signal paire d with the primary. The results of the clinical methods by Johnson et al (2005) incorporated the use of historical data in the statistical analysis. Haller (2006) noted that due to a differing reaction of chickens to the negative antigen protei ns this same calculation system would not be applicable. As was done by Haller (2005) cut off values using ROC curve analysis was performed using 319 known positive and 63 negat ive samples for the WN and 44 known positive and 64 negative samples for the SLE. The raw MFI values were transformed by division of the positive bead set MF I by the negative control antigen for screening and also division of the positive bead se t MFI by the negative antigen bead set for the confirmation. These transformed values were evaluated for the best balance of sensitivity and specificity using Analyse it softwa re (2007). These transformed values, after the MIA test is fully multiplexed, can be run concurrently with the WN and SLE protocols and a database of historical data can the n be accrued and possibly applied to a different calculation scheme in the future. Of the samples that were noted to have different M IA results from their HAI and ELISA counterparts, two sera were found to be poten tially missed positive specimens by the current testing methods. This is a very excitin g finding, it indicates that due to the tests sensitivity and specificity more samples subm itted may be found positive and therefore a more thorough intervention on the behal f of human health can occur. While in

PAGE 100

88 this research only two birds were found positive wi th no previous indication of positivity through HAI testing, it is anticipated that if this test is run weekly on all specimens many more will be potentially found. Both samples came f rom counties that had previously positive birds. Nine sera were found positive in th e MIA that was determined to have come from birds with a positive EEE serum prior to this serum sample. One sample was in fact found positive by MIA four weeks post the o riginal positive HAI result. Three samples older than two weeks (convalescent sera), w hile found positive by the HAI, were negative by the ELISA but positive by the MIA. This illustrates that highly sensitive nature of the MIA assay and the ability to confirm even small levels of IgM in the samples. It is likely that these birds had, in fact proceeded on to secretion of the more specific IgG antibody and that the IgM was in a dec lining phase, and while it was not enough for the ELISA to detect, the MIA still found it possible to quantify. These results in five of the cases were fairly close to the cutof f value for the screening assay, indicating this was a very low level of IgM. In the current testing protocols of the Florida De partment of Health, HAI testing is used as a tool for screening thousands of sera w eekly for IgM or IgG, there is no distinguishing between the two immunoglobulins in t his test. IgM indicates recent infection while IgG indicates a later infection. Th is is of vital importance when determining the risk posed by the disease to humans and livestock in a particular area. The evaluation of the sensitivity, specificity, PPV and NPV against the true value gives a great deal of information about the ability of each test to detect the different antibody types. The true value result is comprised of the co mbined results of several tests. When samples are negative for HAI this sera is considere d to be negative and is not further

PAGE 101

89 evaluated, if in fact the result for the HAI is pos itive, samples are then tested in the MAC –ELISA followed by the PRNT (when necessary), the r esults of each test are evaluated and the final “True Result” is determined. IgM is t he first indicator serologically of an infection, but it can be a week to several weeks be fore the chicken may develop the neutralizing antibodies that will be detected in th e PRNT. At the point where neutralization antibodies come into titers high eno ugh for detection, the IgM antibodies may begin to diminish. The sensitivity of the HAI to the true value indic ated that in the testing for EEE, 99% of the samples that were termed true positives were detected. The specificity of the test was 98% for the EEE virus (Figure 7). The posi tive and negative predictive value of this test was found to be 90% and 99% accurate (Fig ure 8). The sensitivity of this test turned out to be greater than all other testing met hods, however it must be noted that all sera marked negative are not further tested and thi s biased these results. Not all sera were tested by MAC-ELISA or PRNT, but all were tested by HAI. The specificity however proved to be less than all other methods as was the positive predictive value the negative predictive value was on par with the P/N result cal culation but above both the MACELISA and the positive and negative antigen results When comparing the results from the MIA against th e HAI the sensitivity and specificity were found to be 65% and 67% for the sc reening and confirmation respectively, specificity climbs to 99% for both sc reening and confirmation (Figure 9), PPV drops to 88% and 95% and NPV to 95% and 90% res pectively (Figure 10). The drastic decrease is the sensitivity of the MIA vs. the HAI is due to the fact that the MIA is selecting for IgM only, whereas the HAI may be posi tive if either IgM or IgG is present.

PAGE 102

90 In addition the HAI has varying levels of activity that is categorized by the reaction to the varying dilutions of antigen on microtiter plates. These reactions can be labeled “R”,1:10,1:20, or 1:40”, these varying decrees of r eactivity are directly related to the inhibition of the agglutination reaction occurring in each well. This is not an exact measurement as is the quantification by use of the Microsphere assay. It is sensitive to interpretation. It is likely that if stratified by the inhibition titer of the HAI the sensitivity and specificity would show the varying degrees to w hich the true result compares with the varying dilution results. HAI requires great sk ill and experience to complete and properly categorize the result dilution. The MIA re quires no specifically trained person the read the results, only to run the test protocol and plug results into the Excel file. The MAC-ELISA is used as a confirmatory test for t he samples found to be positive in the HAI testing. The basic theory behin d the MAC-ELISA and the MIA are similar. Antibodies in the sera are captured in the ELISA by antigen bound to well of the microtiter plate while the MIA uses a very movable and fluid 5.5-micron bead with carboxylated surfaces upon which many different sub strates can be bound. This flow cytometry based, fluorescent detecting technology a llows even great binding space due to the fact that the bead itself can be bound from any direction all along its surface as apposed to a microtiter plate that only has one sur face exposed for binding of the antibodies holding the antigen. The sensitivity fou nd for the MAC-ELISA vs. the “true result” came to be 80%; this indicates that of all of the negative samples tested in MIA that go on to be tested in PRNT, the ELISA gives th e correct result with 80% accuracy (Figure 7).

PAGE 103

91 The MAC-ELISA is a very labor-intensive IgM only t esting method; its results may also vary due to pipetting error or the use of different reagent lots from one test to another. It is also only used if the sample is show n to be HAI positive at some point, except in the case of this research where 720 sampl es were run regardless of HAI result. The MAC-ELISA test was the highest, however, in spe cificity against the true result at 99% accuracy over all samples tested on the MAC-ELI SA. In other words of all of the samples tested in this research the HAI had a speci ficity of 99% when compared with the true result, again this is testing both IgG and IgM and is therefore not directly comparable to the MIA which tests only IgM. This indicates tha t of those samples tested, it was extremely reliable in indicating the true result th at was elucidated from combination with PRNT and MAC-ELISA. The PPV and NPV against the tru e result were also high at 99% and 90% respectively (Figure 8). A striking finding after calculating the true resu lt vs. testing method statistics was the similarity of the MAC-ELISA sensitivity, specif icity, PPV and NPV to those of the confirmation (Ag+/Ag-) MIA results. Due to the fact that both of these testing methods are based on IgM and both can be considered confirm ation testing methods due to the final calculation of positive antigen divided by ba ckground reaction this was not entirely surprising but was an exciting find. When comparing the MAC-ELISA to the MIA we see tha t the sensitivity of the screening and the confirmation of the MIA is 97% an d 86%, the specificity 95% and 94% (Figure 9), the PPV decreases significantly to 88% and 84% while the NPV increases to 99% and 95% respectively (Figure 10). This is in li ne with the fact that the MAC-ELISA

PAGE 104

92 is also specific for IgM. The Discrepancy between t hese tests may be due to the fact that the MIA has the capability to detect lower levels o f IgM than the MAC-ELISA. The comparison of the “true result” to the P/N MIA result indicated that the sensitivity or 83.6% of the samples were identified correctly by the MIA without additional testing; the specificity on the other ha nd indicates 98.5% were correctly assigned their result value (Figure 7). This is in line with the fact that the true result is based upon both IgM and IgG antibody identification In some cases it was seen that extremely late positives (up to seven weeks post fi rst detection) were not always identified with the screening P/N calculation metho d, this was not unexpected and is probably due to the decrease in IgM over time, they were commonly identified by the confirmation method, but due to the lack of positiv e result with the screening method the overall result was considered negative. The capabil ity of the screening method by itself to positively predict the outcome was at 87%, which wa s less than both the ELISA and confirmation MIA, but above the value of the HAI te st (Figure 7); the NPV was found to be 98% (Figure 8). The confirmation results as previously stated was found to have the same statistical strength as the MAC-ELISA. This testing method assures that it will correctly identify samples potentially positive by screening with 80% sensitivity and a specificity of 99%(Figure 7). Samples that were up to seven wee ks post initial positive were correctly identified with this calculation method i n most cases due to the reactivity to the negative antigen bead set (NSMB protein antigen wit h no viral antigen which then accounted for any reaction in the well to this prot ein instead of the viral antigen). The PPV and NPV were both 92% (Figure 8).

PAGE 105

93 In order to take a closer look at the statistical similarity between the ELISA and the confirmation results, a comparison was done to see if sample identification numbers that were classified as negatives in both tests, bu t that had a true bird result of positive for EEE, were in fact the same sample numbers (Table 14 ). It was hypothesized that the same samples in both sets was the reason the statis tics were so similar. Forty-five samples tested equally tested in both methods were found to be negative with a true bird result of positive for EEE. While all of the sample s analyzed in ELISA and MIA corresponded to each other and were therefore run i n both testing protocols, Twentyseven of the samples were identified as contributin g to the discordance in both testing methods, where forty-nine were found not to be the same in the two tests. In other words an analysis of the discordant results (false positi ves and negatives by the MIA and MACELISA vs. the true result) in the MAC-ELISA and the confirmation of the MIA had only 27 samples in common that contributed to the discor dant results, while 49 of the samples contributing to the discordant values in each testi ng method individually and did not account for the similar statistical result. Finally in order to get a proper look at the actua l Final MIA result, which is considered positive if both the screening and the c onfirmation methods positive but negative if either test gave a negative result, a f inal result was calculated and compared to the true result. The sensitivity and specificity in this final result were also very close to the MAC-ELISA results, the sensitivity was found to be 78% while the specificity was at 99% (Figure 7). The PPV was a strong 98% while the NPV came out at the 93% level, actually slightly higher than the MAC-ELISA at 90% (Figure8). This final result will be the actual result reported after a screening proces s followed by a confirmation test if the

PAGE 106

94 screening test is found to be negative. With histor ical information it may be noted at a future date that the sensitivity of this test is pr obably actually higher than reported in this work, it may be seen to fluctuate through time when factors such a sero-prevalence and submitted sera increase. In the calculation of thes e final results it is possible that some of the old positives (greater than one week post origi nal positive determination) may have been included and it is possible that due to the se nsitivity of this test some of those found positive for EEE by the old method including PRNT ( IgG) were not in fact still positive for IgM (Table 14). Following the analysis of the MIA final result wit h the true result, the MIA was compared against the MAC-ELISA the current gold sta ndard for IgM. The sensitivity and specificity in this case were 84% and 97%. This ind icates a great improvement in the sensitivity of this test if only IgM positives and negatives are compared. The PPV and the NPV were also robust at 92.9 % and 93.6% leading to the conclusion that this test overall is an strong test that once multiplexed will provid e fast and accurate results for the surveillance of the common arboviruses of Florida t hrough the sentinel chicken program. Highlands J is also an alphavirus with cross-react ivity to the antibodies for EEE. Forty samples were tested that were known to be pos itive for antibody to Highlands J virus. Of these, not one serum tested positive by t he screening method. Three did however test positive when using the positive to ne gative antigen bead set (confirmation) transformation method, though again not in the scre ening method and since a final result of positive requires both the screening and confirm ation to be positive the final result by the MIA in these samples is actually negative. This indicates that 100% of the known HJ positive samples were correctly identified and clas sified.

PAGE 107

95 Table 14. Comparison of testing methods to the true value in 2X2 tables. HAI vs. True Value True value HAI EEE Negative Grand Total Negative 2 1970 1972 Positive 242 25 267 Grand Total 244 1995 Elisa vs. True Value True value ELISA Result EEE Negative Grand Total Negative 45 423 468 Positive 191 2 193 Grand Total 236 425 MIA Screening vs. True Value True value Screening result EEE Negative Grand Total Negative 40 1965 2005 Positive 203 30 233 Grand Total 243 1995 MIA Confirmation vs. True Value True value Confirmation EEE Negative Grand Total Negative 45 1000 1045 Positive 191 10 201 Grand Total 236 1010 MIA Final Result vs. True Value True value MIA Final Result EEE Negative Grand Total Negative 47 1007 1054 Positive 170 3 173 Grand Total 217 1010

PAGE 108

96 Figure 7. Comparison of sensitivity and specificity for the Hemagglutination Inhibition test (HAI), MAC-ELISA, and the Microsphere-based Im muno assay (MIA) for the detection of antibodies to Eastern Equine Encephali tis virus (EEE) compared to the true value results (combining both IgM and IgG testing m ethods).

PAGE 109

97 Figure 8. Comparison of Positive and Negative predi ctive values (PPV, NPV) for the Hemagglutination Inhibition test (HAI), MAC-ELISA, and the Microsphere-based Immuno assay (MIA) for the detection of antibodies to Eastern Equine Encephalitis virus (EEE) compared to the true value results (combining both IgM and IgG testing methods)

PAGE 110

98 Figure 9. Comparison of sensitivity and specificity for the Hemagglutination Inhibition test (HAI), MAC-ELISA, against the screening and co nfirmation results of the Microsphere-based Immuno assay (MIA) for the detect ion of antibodies to Eastern Equine Encephalitis virus (EEE).

PAGE 111

99 Figure 10. Comparison of Positive and Negative pred ictive values (PPV, NPV) for the Hemagglutination Inhibition test (HAI), MAC-ELISA, against the screening and confirmation results of the Microsphere-based Immun o assay (MIA) for the detection of antibodies to Eastern Equine Encephalitis virus (EE E).

PAGE 112

100 Figure 11. MIA final results incorporating both the screening and confirmation techniques for one final result vs. the IgM ELISA, the current gold standard for IgM detection in sentinel chicken sera.

PAGE 113

101 Summary and Conclusions The specific aims of this research were to show th at the Luminex Microsphere assay technology can provide an accurate and rapid detection method for IgM antibodies to sentinel chicken sera. Secondly, samples that we re differing in their response to HAI and MAC-ELISA when compared with MIA were analyzed to evaluate if positives were in fact missed by the current methods. Thirdly the sensitivity, specificity, PPV and NPV were evaluated for the HAI, MAC-ELISA and MIA again st the true result as well as in comparison of the HAI and MAC-ELISA to the MIA itse lf. Lastly, to determine if the Highlands J virus cross-reactive antibodies to EEE has any impact upon the testing results. These aims are all vital to show strong s upport for the efficacy of this testing method. The accuracy of the MIA protocol reagents has been shown through individual testing of all reagents and analysis to find the mo st advantageous dilution, thereby increasing accuracy and decreasing cost. This test is in fact more sensitive that the MACELISA and takes a great deal less time. The HAI res ults can take days until a final result is received, the ELISA also takes a period of at le ast two days, the MIA on the other hand takes only hours to screen (thereby limiting the ne ed for HAI to instances where IgG is being sought) and only hours to confirm (thereby li miting the need for ELISA). Samples found to be old positives from convalescent sera of up to seven weeks indicate that the MIA has the ability to correctly classify them with the confirmation screening. While

PAGE 114

102 samples less than five weeks post-infection were fr equestly also seen positive in the screening method the confirmation showed the abilit y to be able to detect them even if the MAC-ELISA could not. This test however does not work on wild bird sera and thus wild birds will still need to be tested using the H AI and PRNT methods. In the future it will be important to generate and store data to inc rease the utility of this testing algorithm. Only through time can data generated by this method produce the historical baseline needed for interpretation of the health im plications of the demonstrated seroconversion rates. The analysis of samples with discrepant results be tween the HAI, ELISA and MIA unearthed two sera that would otherwise have be en labeled negative by the current HAI, MAC-ELISA method. While the sera themselves we re selected to be true negatives the MIA indicates that there is a low level of IgM present (both samples with a P/N of less that twice the cutoff value). The confirmation tranformation however indicates a much stonger reaction both being greater than twice the cutoff value. Future research that may clarify such discrepancies could include the te sting of antibody titers of nave chickens innoculated with the virus drawn at interv al periods to assess the strength of the response in time in both the HAI, MAC-ELISA as well as the MIA. In that case the birds would have a known innoculation of the virus and ev aluation can be made if in fact the HAI and MAC-ELISA are indeed missing potential posi tives as the MIA suggests. Another suggestion for further research includes th e analysis of results in the MIA and MAC-ELISA by stratification of the inhibition tire of the HAI. Using this technique evaluation could be made of the result received in confirmatory testing in relation to the strength of the titer reported (“R” vs. 1:40).

PAGE 115

103 Alphaviruses are known to be extensively cross-rea ctive. Highlands J commonly co-circulates through the bird population with EEE. This alphavirus however is not a pathogen to humans, it causes no disease process of which we are currently aware. However it can cause a positive result for alphavir us antibodies in the HAI method. It is not possible to delineate between a bird infected w ith HJ or EEE based solely on this testing method; MAC-ELISA does however have the abi lity to differentiate, for it tests only for EEE reactivity. MIA screening methods corr ectly classified 100% of the HJ positive sera to be negative for EEE. While the con firmation calculations alone showed three positive results, these sera were determined to have a final MIA result of Negative due to the fact that the screening protocol showed no reactivity and a positive in both the screening as well as the confirmation method is nec essary for a final result of positive. It is clear that when this protocol is correctly imple mented with the screening method first there should be no reason cross-reactivity will eff ect the result. Overall the implementation of the Microsphere assa y for Eastern Equine encephalitis in sentinel chickens will save the lab oratories both time and money in the quest for results and infromation to archive and ev aluate. Future projects should include the multiplexing of this assay with the WN and SLE protocol designed by Logan Haller (2006) and evaluation of data archived after implem entation of this testing method in a high throughput situation. This data can be evaulat ed toward the goal of more difined cutoff values or re-evaluation of the data transfor mation methods.

PAGE 116

1 References Cited Abbas, A.K. and Lichtman, A.H. (2005). Cellular and Molecular Immunology (5th ed.) Philadelphia: Elsevier Saunders. Alan L. Schmaljohn and David McClain (1996).Virolog y: Alphaviruses. In Baron Samuel, M.D., ed. Medical Microbiology ( 4th ed.) The University of Texas Medical Branch at Gal veston. APHIS veterinary services Fact Sheet; Eastern Equin e Encephalitis. (2003) Arboviruses. (1999, February 16). Retrieved Decembe r 15, 2007, from .html Bell, J.R., Kinney, R.M., Trent, D.W., Strauss, E.G ., and Strauss, J.H. (1984). An Evolutionary Tree relating eight Alphaviruses, based on amino-termina l sequences of their glycoproteins. Proc. Natl. Acad. Sci., 81:4702-4706. BioRad Laboratories. (2005) BioPlex Manager Softwar e 4.1 User guide. p187Byrne R.J. and Robbins, M.L. (1960). Mortality patt erns and antibody response in chicks inoculated wit h eastern equine encephalomyelitis virus. Journal of Immunology 86:13-16. Calisher, C.H. (1994). Medically Important Arboviru ses of the United States and Canada. Clinical Microbiology Reviews 7(1), 89-116. Calisher, C.H., Fremount, H.N., Vesley, W.L., El-Ka frawi, A.O., and Al-Deen Mahmud, M.I. (1986). Relevance of Detection of Immunoglobulin M antibody Response in Birds Used for Arbovirus Surveillance. Journal of Clinical Microbiology 24(5):770-774. Centers for Disease Control and Prevention (2007).W est Nile Virus Transmission Through Blood Transfusion South Dokota 2006. MMWR Morb Mortal Wkly Rep 56(04):76-79. Centers for Disease control and Prevention (2004). West Nile Virus: Information and guidance for clinicians. e/clinicians/pdf/wnv-clinicaldescription.pdf. Centers for Disease Control and Prevention (2005). Eastern Equine Encephalitis – New Hampshire and Massachusetts, August – September 2005. MMWR Morb Mortal Wkly Rep55(25):697-700. Chamberlain R.W. and Sudia, W.D. (1961). Mechanism of Transmission of Viruses by Mosquitoes. Annu. Rev. Entomol 6:371-390. Chonmaitree, T., Baldwin, C.D., and Lucia, H.L. (19 89). Role of the Virology Laboratory in Diagnosis a nd Management of Patients with Central Nervous System. Clinical Microbiology Reviews 2(1):1-14. Davis, E. (2005) Cause and Prevention: vector-media ted encephalitis: vaccinations, insect management c an help stifle viral impact. DVM Newsmagazine. 36(3):1-8. Day, Jonathan F., & Stark, L.M. (1996). Transmissio n Patterns of St. Louis Encephalitis and Eastern Equine Encephalitis Viruses in Florida: 1978-1993. Journal of Medical Entomology 33 132-139. Evans, Borton, Hintz, Van Vleck (1977). The Horse. W.H. Freeman and Company. New York. Farrar, M.D., Miller, D.L., Baldwin, C.A., Stiver, S.L., and Hall, C.L. (2005). Eastern Equine Encepha litis in dogs. J. Vet. Diagn. Invest. 17:614-617.

PAGE 117

2 Franklin, R.P., Kinde, H., Jay, M.T., Kramer, L.D., Green, E.N., Chiles, R.E., Ostlund, E. Husted, S ., Smith, J., and Parker, M.D. (2002). Esatern Equine Encephalomyelitis virus infection in a horse from California. Emerging Infectious Diseases 8: 283-288. Gould, E.A., Higgs, S., Buckley, A. and Gritsun, T. S. (2006). Potential Arbovirus Emergence and Implications for the United Kingdom. Emerging Infectious Diseases 12(4): 549-554. Guy, J.S. and Malkinson M. (2003). Section I Viral Diseases: Arbovirus Infections. In Y.M. Saif (Ed.), Diseases of Poultry. Blackwell Publishing. Helm, J.D. (2003). EEE and Emus and Pheasants. Clem son University, Livestock Poultry Health Programs. Heshati, H. (2004). Florida Department of Health Br evard County – Advisory for West Nile Virus. Tallahassee, FL. Holden, P., Muth, D. and Shriner, R.B.(1966). Arbov irus Hemagglutination-Inhibition in Avian Sera: Inactivation with Protamine Sulfate. American Journal of Epidemiology 84(1):67-74. ICTVdB Management (2006). 00.026. Flaviviridae. In: ICTVdB The Universal Virus Database version 3. Bchen-Osmond, C. (Ed), Columbia University, New Yo rk, USA Jacob, J. (2003). Viral Encephalitis: The role of b irds. IFAS extension: University of Florida Factshe et PS50. Janeway, C.A., Travers, P., Walport, M., Shlomchik, M.J. (2005). Immunobiology (5th ed.) New York: Garland Science Johnson, A.J., Cheshier, R.C., Cosentino, G., Masri H.P., Mock, V., Oesterle, R., Lanciotti, R.S., Ma rtin, D.A., Panella, A.J., Kosoy, O., and Biggerstaff, B. J. (2007). Validation of a Microsphere-based Immunoassay for Detection of an anti-West Nile Viru s and anti-St. Louis Encephalitis Virus Immunoglobulin M antibodies. Clin. Vaccine Immunol. 14(9):1084-1093. Johnson, A.J., Noga, A.J., Kosoy, O., Lanciotti, R. S., Johnson, A.A., and Biggerstaff, B.J. (2005). Du plex Microsphere-based Immunoassay for the detection of anti-West Nile Virus and anti-St. Louis Encephalitis Virus Immunoglobulin M antibodies. Cli nical and Diagnostic Laboratory Immunology 12(5):566-574. Kellar KL and Iannone MA (2002). Multiplexed micros phere-based flow cytometric Assays. Experimental Hematology 30:1227-1237. Kelso, K. and Trahan, B. (1999) Infectious Disease Epidemiology Annual Report 1999. Louisiana Department of Health and Hospitals Office of Public Health. Kissling R.E. (1960). The Arthropod-borne viruses o f Man and other Animals. Annu. Rev. Microbiol. 14:261-282. Kuno, G. and Chang G.J. (2005). Biological Transmis sion of Arboviruses: Reexamination of and New Insight into Components, Mechanisms, and Unique Tra its as well as Their Evolutionary Trends. Clinical Microbiology Reviews, 18(4):608-637. McLean RG, Crans WJ, Caccamise DF, McNelly J, Kirk LJ, Mitchell CJ, Calisher CH. (1995) Experimental infection of wading birds with eastern equine encephalitis virus. J Wildl Dis 31:502– 508.

PAGE 118

3 McLintock, J. (1978) Mosquito-virus Relationships o f American Encephalitides. Ann. Rev. Entomol., 2317-37. Moncayo, A.C., Edman, J. D. and Turrell, M.J. (2000 ). Effect of Eastern Equine Encephalitis virus on t he survival of Aedes albopictus, Anopheles quadrimacul atus and Coquilletidia perturbans. J. Med. Entomol 37(5): 701-706. Moore, C.G., McLean, R.G., Mitchell, C.J., Nasci, R .S., Tsai, T.F., Calisher, C.H., Marfin, A.A., Moor e, P.S., and Gubler, D.J. (1993). Guidelines for Arbov irus surveillance programs in the United States. Center for Disease Ceontrol and Prevention: Division of Vector-Borne Infectious Diseases. Nandular, M. and Urban A.W.(2007). Eastern Equine E ncephalitis. Retrieved February 25t, 2008, from emedicine from WebMD. topic3155.htm Olleveant, N.A., Humphris, G., Roe, B. (1999) How B ig is a Drop? A volumetric assay of essential oils. J. of Clin. Nursing ( 8):299-304. Paessler, S., Aguilar, P., Anishchenko, M., Wang, H ., Aronson, J., Campbell, G., Cararra, A., and Weav er, S. (2004) The Hamster as an Animal Model for Easter n Equine Encephalitis and its Use in Studies of Virus Entrance into the Brain. Journal of infectious Disease 189:2072-2076. Passler, S. and Pfeffer, M. (2003). Detection of An tibodies to Alphaviruses and Discrimination between Antibodies to Eastern and Western Encephalitis Viru ses in Rabbit Sera using a Recombinant Antigen and Virus-specific Monoclonal Antibodies. J. Vet. Med 50: 265-269. Pier, G.B., Lyczak, J.B. and Wetzler, L.M. (2004). Immunology, Infection, and Immunity. Washington D.C.: ASM press. Poonacha, K.B., Gregory, C.R., and Vickers, M.L. (1 998). Intestinal lesions in a horse associated with eastern equine encephalitis virus infection. Vet. Pathol 35:535-538. Powers, A.M., Brault, A.C., Shirako, Y, Strauss, E. G., Kang, W, Strauss, J.H., & Weaver, S.C. (2001). Evolutionary Relationship and Systemics of the Alph aviruses. Journal of Virology. 75 1011810131. Service, M. W. (2004). Medical Entomology for Stude nts. Third Ed. (pp. 65-70) New York, New York: Cambridge University Press. Smith, A.L. and Anderson, C.R. (1980). Susceptibili ty of two turtle species to eastern equine encephal itis. Journal of Wildlife Disease 16(4):615-617. Unnasch, Robert S., Sprenger, Tonya, Katholi, Charl es R., Cupp, Eddie W., Hill, Geoffrey E., and Unnasch, Tjomas R. (2006). A dynamic transmission m odel of eastern equine encephalitis virus. Ecol Modell. 192, 425-440. Viertlboeck, B.C, Schweinsberg, S., Hanczaruk, M.A. Schmitt, R., Du Pasquier, L., Herberg, F.W. and Go¨ bel, T.W. (2007). The chicken leukocyte recepto r complex encodes a primordial, activating, high-affinity IgY Fc receptor. PNAS 104(28):11718-11723. Weaver, S.C., Kang, W., Shirako, Y., Rumenapf, T., Strauss, E.G., and Strauss, J.H. (1997). Recombinational History of Molecular evolution of W estern Equine Encephalomyelitis Complex Alphaviruses. Journal of Virology 71(1):613-623.

PAGE 119

4 Wellings, F.M., Lewis, A.L., and Pierce, L.V. (1972 ) Agents encountered during arboviral ecological studies: Tampa Bay Areas, Florida, 1963 to 1970, A merican Journal of Tropical Medicine and Hygiene 21(2):201-213.

PAGE 120

Bibliography /cchf.htm cations.htm?seq_no_115=160736 html

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002000377
003 fts
005 20090421085632.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 090421s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002480
RA425 (Online)
1 100
Fitzpatrick, Kelly Ann.
0 245
Luminex microsphere immunoassay offers an improved method in testing for antibodies to Eastern Equine Encephalitis virus in sentinel chicken sera
h [electronic resource] /
by Kelly Ann Fitzpatrick.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 103, 4 pages.
Thesis (M.S.P.H.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Eastern Equine Encephalitis virus has a human mortality rate of 30% of those cases diagnosed, while 30% of those surviving infection remain with neurological sequelae for life (, 2007). The use of sentinel chickens for surveillance of arboviruses that are known to use birds as a reservoir host, such as St. Louis Encephalitis (SLE), West Nile (WN) virus, Eastern Equine Encephalitis (EEE) and Highlands J (HJ) virus, in Florida began with the Sentinel Chicken Arboviral Surveillance Network in 1978 (Day and Stark, 1996). This network enables the activation of an early warning system for citizens, as well as, county epidemiologists and those in mosquito control, allowing for a coordinated effort of disease prevention.Methods currently used at the Florida Department of Health, Tampa Branch Laboratory include screening of submitted sera for antibodies to these arboviruses of epidemiologic importance by way of the hemagglutination inhibition test (HAI), and confirmation by the IgM antibody capture enzyme-linked immunosorbent assay (MAC-ELISA) and Plaque Reduction Neutralization test if the MAC-ELISA proves to be negative. While these tests combined are providing the results needed, the time to result can be a week or greater depending on the initial screening result in the HAI tests. The Microsphere Assay Technology provides an accurate, more rapid (a day or two instead of a week or more) detection method including both a screening and confirmation protocol specifically designed to test for antibody to EEE in sentinel chicken sera.Two sera out of the thousands tested that were tested by HAI shown to be negative in standard testing, resulted as positive by the MIA method and therefore indicated a missed positive. The sensitivity and specificity, positive and negative predictive values of this new protocol as compared with MAC-ELISA as a reference standard indicated that both tests were remarkably similar; Providing sensitivity near 80%, specificity and PPV at 99%, and negative predictive values at 90% for MAC-ELISA and 94% for the MIA. Finally it was determined that Highlands J virus will not have any impact on the testing protocol and results of this test.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Lillian M. Stark Ph.D.
Dissertations, Academic
x Global Health
t USF Electronic Theses and Dissertations.
4 856