The Immune response in horses to vaccination against equine influenza virus. A comparison of recombinant DNA, inactivate...


Material Information

The Immune response in horses to vaccination against equine influenza virus. A comparison of recombinant DNA, inactivated virus, and modified live virus vaccines
Physical Description:
ix, 167 leaves : ill. ; 29 cm.
Sweat, James Mark, 1960-
Publication Date:


Subjects / Keywords:
Horses   ( mesh )
Orthomyxoviridae -- immunology   ( mesh )
Vaccines, Synthetic   ( mesh )
Vaccines, Inactivated   ( mesh )
Vaccines, Attenuated   ( mesh )
Influenza -- immunology   ( mesh )
Influenza -- prevention & control   ( mesh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 2001.
Includes bibliographical references (leaves 150-165).
Statement of Responsibility:
by James Mark Sweat.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 52305413
System ID:

Full Text








Portions of this work were supported by an ongoing project within our laboratory funded

by Hoechst Roussel Vet. I would like to thank Norbert Kline and his colleagues. I would also

like to thank Mark Abdy for his help with the horse work and introducing me to Paul Gibbs. I

would like to give credit to members of my committee, Bradley Bender, Jorge Hernandez, Calvin

Johnson, Paul Lunn, and Parker Small for their guidance and suggestions through the

development of this project. Individuals that were not on my committee but were no less helpful

include, Dorothy Guantlett, Carlos Romeros, Steeve Giguere, Cynda Crawford, and Alysia Posey.

I would like to give special thanks to my committee chair, Paul Gibbs for his wisdom and support

in the development and completion of this project.

I was truly fortunate to have the support and prayer from friends and family. Thanks to

Andy Albertson and Rick Mercer for a constant supply of encouraging words. Thanks to my

parents Boyce and Marceil Sweat for their thoughts and prayers.

Possibly the greatest motivating force was the love and support from my wife Susan and

daughter Jessica. I thank them for their sacrifices and patience throughout my absence.


ACKNOWLEDGMENTS........................................ ..................................... ii

KEY TO ABBREVIATIONS......................................................... .....

ABSTRACT.......................................................... ............... ............. ii



Introduction ................................................................ ..........................

Equine Influenza Virus................................................................................... 2
Structure Design......................................................... ......... ...3
Structural Proteins and Genes.........................................................4
Viral Transmission .................................... .............. ........... .....9
RNA Replication and Protein Synthesis........................................... 10

The Significance of Equine Influenza..................................................... ...........
Host Specificity.........................................................................12
Shift and Drift...........................................................................14
Mode of Transmission and Infection................................................. 15
Clinical Presentation................................................................... .... 16
Pathogenesis........................................................................... ...17
Epidemiology................................................................. .... 18
Impact to the Industry................................................................. 19

Vaccination Against Equine Influenza Infection.................................................... 21
Vaccine Theory.........................................................................21
Killed Vaccines........................................................................23
Live Vaccines.......................................................................... 23
Recombinant DNA Vectors...........................................................25
Immune Response.......................................................................25
Antigen Presentation....................................................................26
Duration of Antibody Post-Infection
or Vaccination.................................................................27
Mucosal Immunity......................................................................28
Hypothesis and Objectives................................................................................30


H hypothesis ................................................ ..........................................30
O bjectives...........................................................................................30


Introduction ............................................................................................... 31

Materials and Methods............... .................................................. .................... 33

H I A ssay ................................................................................... ........ 33
SRH Assay............................................ ......... ............................. 34
ELISA Standardization Techniques.................. .................................. ... 34
Virus purification for use in ELISA ................................................. 34
ELISA development..................................................... ........... 35
Sensitivity and Specificity.................................................................... 37
Lymphocyte Proliferation Assay................................................................38
Cytokine mRNA Determination in Equine PBMC......................................... 40
RNA isolation and reverse transcription............................................40
Realtime PCR analysis .................................................. ............ 42

Determination of the Critical or Cut off Value for the ELISA............................. 46
Generating a "Critical Value" Plot........................................................... 47
Lymphocyte Proliferation Assay ................................................. ........ ....48
Cytokine mRNA Determination in Cryopreserved Equine PBMC.........................56

D discussion ................... .............................................. ................................ 64


Introduction...................................................................... .................... 72

Materials and Methods....................................................................................... 73
H orses ..................................................................................... 73
Infection Strains of Equine Influenza Virus ...................................................73
Infection Model Validation Study..............................................................74
Nasal Secretion Sampling........................................... ............. ............ 77
Absorption of Antibody by Rayon Swabs.................................... .............. 78

R esults............................................................... ........................................78
Validation of Infection of Horses with Equine Influenza Virus........................... 78
Sampling Method Used to Collect Nasal Secretions............................. .............81
Determination of Antibody Resorption by Rayon-Tipped Swabs.......................... 83



Introduction................................................................................................. 85

Materials and Methods............... ..................................................................86
Horses........................ .................... .......... ............... ........... 86
Vaccines................................................. ....................................87
Virus............................................................... ... .......... ...................90
Sample Collection and Storage.................................................. ..............90
Serum..................................................................... .........90
Peripheral blood mononuclear cells...................................................90
Nasal secretions for IgA antibody determination.................................91
Experimental and challenge infection................................................91
Virus isolation...........................................................................91
Clinical Signs ..................................................................................92
Virus Isolation and Titration..................................................................93
Nasal Antibody Determination................................................................. 94
Cell-Mediated Immunity......................................................... ........... 94
Study Timeline....................................................................................... 94
Statistical Analysis of the Immune Response in Vaccinated
and Infected Horses........... .......... .............................................95

Results............................................................................. .......................96
Clinical Scores ...................... ........................................ ...... ..... 96
Virus Shedding in Vaccinated and Infected Horses..........................................98
SRH ................................................................................ .... ....... 106
ELISA............................................................................................. 108
Local IgA Antibody Response..................................................................110
Cell-Mediated Immunity......................................................................... 12
Summary of Immune Responses............................................................. 115
Cytokine Determination..........................................................................122


Future Studies.................................................................................................146

5 GENERAL SUMMARY................................................................................148

REFERENCE LIST.........................................................................................150

BIOGRAPHICAL SKETCH............................................................................166


ANOVA Analysis of variance
APC Antigen presenting cells
C Centigrade
CM Culture medium
CMI Cell-mediated immunity
Con A Concanavalin A
CS Clinical score
CTL Cytotoxic T-lymphocyte
DMSO Dimethyl sulfoxide
DTT Dithiotreatole
EHV-4 Equine herpes virus-4
EID Egg infectious dose
EIV Equine influenza virus
ELISA Enzyme-linked immunosorbant assay
ER Endoplasmic reticulum
G3PDH Glyceraldehyde-3 phosphate dehydrogenase
gE Glycoprotein E
gI Glycoprotein I
HA Hemagglutinin
HAU Hemagglutinating unit
HCL Hydrochloric acid
HI Hemagglutination inhibition
hrIL-2 Human recombinant interleukin-2
IACUC Institutional Animal Care and Use Committee
IgA Immunoglobulin A
IgG Immunoglobulin G
IgM immunoglobulin M
IL-4 Interleukin-4
IL-6 Interleukin-6
IM intramuscular
IN Intranasal
INF-y Interferon gamma
ISCOMS Immune stimulating complexes
IURD Infectious upper respiratory tract disease
M1&2 Matrix protein
MALT Mucosal-surface-associated lymphoid tissue
MDCK Madin Darby canine kidney
MHC Major histocompatibility complex
MLV Modified-live virus
MMLV Moloney murine leukemia virus
NA Neuraminidase
NALT Nasal-associated lymphoid tissue

NP Nucleocapsid protein
NS Nonstructural
NVSL National Veterinary Services Laboratory
OD Optical density
OPD O-phenylenediamine dihydrochloride
PBMC Peripheral blood mononuclear cells
PHA Phytohemagglutinin
RBC red blood cell
Recomb Recombinant
RNP Ribonucleoprotein
rtPCR Reverse transcription polymerase chain reaction
SC Secretory component
SI Stimulation index
SRH Single radial hemolysis
TCID Tissue culture infectious dose
TCR T-cell receptor
TGN trans Golgi network
Thl and 2 T-helper
TNF-a Tissue necrosis factor-alpha
UK United Kingdom
URD Upper respiratory tract disease
UV Ultraviolet
VTM Virus transport medium
WHO World Health Organization

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirement for the Degree of Doctor of Philosophy



James Mark Sweat

August, 2001

Chair: Dr. Paul Gibbs
Major Department: Veterinary Medicine

Equine influenza virus is a cause of upper respiratory disease that has economic

significance to the horse industry. It is a highly contagious virus that is spread among horses

despite vaccination. Due to antigenic changes in surface proteins and the ineffectiveness of

vaccines, equine influenza has been maintained as an endemic disease within the equine


Various studies have shown that the use of inactivated-virus vaccines results in short-

lived immunity that does not prevent infection but reduces the incidence of severe clinical

disease. The implications of this partial immunity in equine influenza are two-fold. A sub-

clinical infection will often not only show a reduction in performance, but serve as a potential

threat to susceptible horses during racemeets and show events.

Because conventional inactivated vaccines against equine influenza are ineffective, novel

approaches using recombinant DNA technology have been investigated. This study examines the

potency of a recombinant DNA vaccine consisting of an equine herpes-4 virus vector expressing

the hemagglutinin and neuraminidase genes of the A/equine/2/Kentucky/94 strain of equine

influenza virus. The immune response in horses was determined following inoculation with a

recombinant DNA vaccine and compared to that seen after using a modified-live and inactivated-

virus vaccine or natural infection. Horses were then challenged with equine influenza virus and

the ability of each vaccine to prevent infection or reduce severe clinical disease was determined.

The results of this study indicate that the recombinant DNA vaccine examined here failed

to induce an immune response due to over-attenuation. The modified-live and inactivated virus

vaccines provided a reduction in the severity of clinical disease compared to that of an un-

vaccinated control. Protection from severe disease was associated with serum antibodies in the

inactivated vaccine recipients and serum, local, and cell-mediated immunity in the modified-live

vaccinates. Horses initially infected with live virus were completely protected against infection

upon challenge 8 weeks later. This study supports previous investigations and concludes that 1)

inactivated-virus vaccines provide incomplete protection through humoral immunity, 2)

modified-live vaccines induce serum and local antibody formation, and a cell-mediated immune

response, and 3) repeat vaccinations are required to maintain immunity against infection with

equine influenza virus.



Equine influenza is an infectious upper respiratory disease of horses that has economic

significance. Equine influenza has been the cause of outbreaks that have had an effect on race-

meets, show and performance events, and breeding programs on an international level. Because

of this, vaccines against equine influenza virus (EIV) infection have been investigated for the past

40 years. In 1956, EIV was first isolated in Prague, Czechoslovakia, and recognized as being a

cause of upper respiratory disease in horses (Sovinaova, Tumova, and Pouska 1958). In 1963, a

new strain was reported from horses at a racetrack in Florida and designated

A/equine/2/Miami/63 (H3N8) (Wadell, Teigland, and Sigel 1963). Since the appearance of the

type-2 strain of EIV, a number of sub-types have emerged leading to increased monitoring efforts

and new strategies in vaccine development.

Due to the inability of conventional inactivated-virus vaccines to provide complete

protection against infection and disease, novel vaccines such as recombinant DNA and modified-

live preparations have been investigated. Studies including adjuvant-conjugated inactivated-

virus (Hannant, Easeman et al., 1999;Mumford, Jessett et al., 1994a;Mumford, Wilson et al.,

1994), modified-live (Holland, Chambers et al., 1999); (Lunn, Hussey et al., 2001), and DNA

vaccines (Lunn, Soboll et al., 1999) have been conducted to determine the level of immunity and

protection induced by each. Historically, the primary surrogate of immunity to EIV was an

increase in antibodies to the hemagglutinin protein (Morley, Hanson et al., 1995). More recent

studies have included cell-mediated effector mechanisms (Hannant, 1994;Hannant & Mumford,

1989;Hannant, 1994;Morley, Hanson, Bogdan, Townsend, Appleton, & Haines, 1995;Mumford

& Wood, 1992). As the availability of equine-specific reagents increased, so did studies that

characterize sub-isotype-specific antibody formation (Lunn, Olsen et al., 1999;Lunn, Hussey,

Sebing, Rushlow, Radecki, Whitaker-Dowling, Youngner, Chambers, Holland, Jr., & Horohov,

2001;Nelson, Schram et al., 1998) and cytokine expression in horses (Giguere & Prescott, 1999).

Investigations that examine a comprehensive immune response following vaccination or infection

are still warranted, however. In this study, the immunological response of horses to vaccination

against or infection with equine influenza was measured. To accomplish this, traditional and

novel assay methods to quantify the immune response were optimized. Furthermore, the

optimized methods were then used during animal studies that included immunization of horses

with recombinant DNA, inactivated-virus, or modified-live vaccines.

Equine Influenza Virus

Equine influenza virus is one of a group of enveloped viruses in the family of

Orthomyxoviridae which in the Greek translation means, ortho, "standard or correct," and myxo,

"mucous." The name "influenza" can be traced to a Latin origin influential, or influence,

because it was once believed that the cause of this disease had astrological or occult influences

(Murphy, Gibbs et al., 1999). The current nomenclature was determined by the International

Committee on Taxonomy of Viruses (ICTV) in the 1960s. To reference a particular strain, the

system includes host origin, geographical location of the first isolate, strain number, and year it

was isolated (WHO Memorandum, 1980). Orthomyxoviruses have a segmented, single stranded

RNA, negative-strand genome. Because they are negative strand, they must not only contain a

template for mRNA synthesis but also for anti-genomic (+) strand RNA (Fields, 1996).

Orthomyxo viruses are similar to another group of viruses, the Paramyxoviridae, which have

similar biological and structural properties. Influenza viruses use a number of unique and

complex mechanisms for the replication of viral proteins. Viral replication includes not only host

protein-synthesis-machinery but host-cell RNA for primers as well. This mode of replication is a

unique example of RNA that is not transcribed from DNA by RNA polymerase II.

Equine influenza virus is a type A virus. Influenza viruses, types A, B, and C, infect

humans, but are of little importance in veterinary medicine and none have been recorded to cause

clinical infection in the horse. The various types and strains of influenza viruses are separated

from each other based on antigenic specificity resulting from structural differences. Some other

unique properties of the influenza viruses are listed in the table below:

TABLE 1-1. Major Characteristics of Influenza Viruses

Type Unique Properties
A viruses Infect mammals including humans, swine and horses
Infect avian species
Surface proteins have a greater diversity as compared to B and C
Contain eight distinct RNA segments (As does B)

B Infect only humans
Less complex surface glycoprotein amino acid sequence
Contain eight distinct RNA segments (as does A)

C Isolated from humans primarily but are also seen in swine
Have a single multifunctional glycoprotein
Contain 7 RNA segments

Adapted From Fields (1996)

Structural Design

The structure of the influenza virus varies slightly among types and has basic components

made up of 0.08% to 1% RNA, 70% protein, 20% lipid, and 5% to 8% carbohydrate (Fields,

1996). Because influenza virus exits the infected cell through a process called 'budding", the

lipid envelope is essentially made up of the plasma membrane from the host cell. Influenza virus

is morphologically recognizable resulting from the presence of the numerous (approximately 500,

10-14 nm) hemagglutinin (HA) spikes protruding from the lipid envelope (Laver & Valentine,

1969). Shorter in length, the mushroom shaped neuraminidase (NA) spikes are dispersed among

the HA proteins. The ratio of these two membrane proteins is 5:1 HA to NA respectively

(Zebedee & Lamb, 1988). Also present in the membrane is the viral matrix protein (MI) and the

transmembrane protein M2 ion channel. Within the core of the viral structure are

ribonucleoprotein (RNP) and nonstructural proteins (NS1 and NS2). The RNP structures consist

of 8 separate segments with unique genomic products. Of particular importance in terms of

antigenic recognition is segment 4 which encodes the HA protein, the major surface-glycoprotein

sialic acid-binding structure. Other segments are responsible for the synthesis of P (polymerase)

proteins (PB1, PB 2, and PA), and nucleocapsid (NP) (the transcriptase complex) as well as NA,

M1, M2 and NSI and 2 (Lamb & Choppin, 1976).

Structural Proteins and Genes

Polymerase proteins. PB2 and PA are responsible for encoding the largest of the viral proteins

that are aggregated in the cytoplasm and nucleus of infected cells. The PBI protein has a key role

in catalyzing nucleotide polymerization and chain elongation (Braam, Ulmanen et al., 1983).

Nucleocapsid proteins. The NP reacts with other RNA segments to form a nucleocapsid. The

NP protein has been the subject of research to determine homology between different strains.

Cytotoxic T lymphocytes against influenza A virus will non-specifically cross-react with NP

subtypes of influenza A virus isolated from humans and mice (Yewdell, Webster et al., 1979).

The NP is synthesized in the cytoplasm, complete with specific residues that provide targeting to

the nucleus for association with the various RNA strands for assembly (Lin & Lai, 1983).

Hemagglutinin proteins. The HA has a molecular weight of 61,468 with 1,778 nucleotides; it

binds to the sialic acid glycoprotein receptors on erythrocytes and other cells, and is responsible

for attachment and penetration into the host cell (Fields, 1996). There is often considerable

antigenic variation between the HA protein of different strains of influenza virus and thus the HA

has a significant role in "defining" the susceptible host. The HA protein mediates fusion of the

virus membrane and the endosomal membrane facilitating the release of the nucleocapsid

contents into the cytoplasm of the host cell (Fields, 1996). Structurally, the HA starts as a

homotrimer manufactured in the endoplasmic reticulum (ER) designated Ho (Wiley & Skehel,

1977). After modifications of amino acid residues added to the carboxy-terminal, the HA

molecule is targeted to the ER membrane by an N-terminal signal peptide. The integral

membrane protein then consists of two segments with a di-sulfide linked region (Fields, 1996).

The H, segment consists of 319 to 326 residues and is cleaved leaving an H2 segment

(Verhoeyen, Fang et al., 1980). The cleavage site can consist of up to 6 residues depending on

the subtype and shows the loss of a single arginine residue. The loss of an arginine residue

would indicate a trypsin-like or carboxypeptidase B type enzyme which is responsible for the

activation of the HA protein during attachment (Dopheide & Ward, 1978). During infection,

changes occur to the HA molecule within endosomes. HA molecules within the low pH

environment of the endosome are cleaved, exposing a fusion peptide. Depending on the virus

strain, the type of cells infected, and growth conditions, the HA can be whole (HA0) or cleaved

into HAI and HA2. The HA protein must be cleaved into HAI and HA2 for the virus to be

infectious (Lamb, 1989). The fusion peptide requires the cleavage of HAo into HA, and HA2 for

activation. This activity has a specific temporal sequence which must be exact to avoid the

formation of oligomer aggregates. Cleavage of the connecting peptide of the HA molecule occurs

in response to different compounds. Furin is an intracellular enzyme found within the trans Golgi

network which will cleave HA (Stieneke-Grober, Vey et al., 1992). Hemagglutinin from virus

grown in cell-culture systems requires a source of exogenous trypsin if the HA contains an

arginine residue in the connecting peptide; this is seen in all human strains. When influenza is

cultured in embryonated eggs, a different enzyme proposed to be a factor Xa-like protease is

responsible for cleavage of the HA (Kawaoka & Webster, 1988). Virulence versus non-

virulence is thought to be related to the presence of furin recognition motifs or a single arginine

residue respectively (Nestorowicz, Kawaoka et al., 1987).

As previously mentioned, the HA is a major component of antigenic recognition by

neutralizing antibodies. The epitopes of the HA are made of two regions; one extends

approximately half way up the trimer (HA2 region) and a second at the globular head of the

molecule (HAI). These epitopes form the HA receptor binding site at the end of each HAI and

HA2 sub-unit (Gibson, Daniels et al., 1988). Antigenic variations are demonstrated by changes or

substitutions at specific amino acid residue sites ((Rogers & Paulson, 1983). Antigenic

relatedness has been established by comparing the homology between different segments of the

HA protein. By comparison, nucleotide and amino acid substitutions are implicated in low levels

of antigenic drift. For example, 16 subtypes of type 1 equine influenza isolated during a period

ranging from 1956 to 1977 were compared based on amino acids deduced from nucleotide

sequences of viral RNA for the HA protein. Additionally, the genomic sequence of the HA from

two other influenza viruses from avian origin (A/FVP/Rostock/34 (H7N7) (Porter, Barber et al.,

1979) and seal origin (A/seal/Massechusetts/1/80 (H7N7) were compared (Webster, Hinshaw et

al., 1981). When the HA gene nucleotide sequence of A/equine/Prague/53 subtype was

compared to the Rostock avian and Massachusetts seal subtypes, the signal peptide, the HAl, and

the connecting peptide, consisting of 1,068 bases, showed homologies ranging from 55 to 77

percent. When amino acid extensions at the carboxy-terminus were compared among this group,

an even higher percentage of homology was noted ranging from 79% to 84%. This study

showed that a high degree of homology was present between strains that were considered to share

an ancestral relationship. In contrast, another study comparing avian and equine species of

influenza revealed no carboxy-terminus extensions in the avian subtypes which are characteristic

in equine influenza (Fields, 1996). A similar study comparing equine subtypes Prague/56 and

Cambridge/63 revealed a high degree of similarity with only 6 nucleotide differences in the HAl

coding sequence (Webster, 1993). Five such nucleotide changes resulted in amino acid

substitutions. Of particular interest was the change of Thr to Ile at site 160 in the Cambridge/63

virus; a potential glycosylation site. Additions and deletions of amino acids at these potential

glycosylation sites are considered to be relevant when comparing the relatedness of strains but,

more importantly, are thought to be associated with virulence. These areas are regarded as

important functional (fusion properties) sites based on their three-dimensional structures

(Daniels, Skehel et al., 1985).

Modifications and conservation in amino acid sequences are implicated in affecting the

infectivity of an emerging subtype. When the A/Prague/63 HAl sequence was aligned with a

human influenza A (H3N2), cysteine residues were found to be important in maintaining the

three-dimensional structure as well as other various residues (Daniels, Skehel, & Wiley,

1985;Rogers & Paulson, 1983). Additionally, all strains of EIV-1 have a conserved glutamine at

site 226 which has significant binding specificity with the alpha-2:3 sialic acid receptor. This

indicates a conserved function within the HA molecule of different strains.

The rate at which both silent mutations and amino acid substitutions occur is believed to

be a result of the frequency of replication cycles and can have significance in strain diversity

(Fields, 1996). The rate is highest in human influenza followed by swine and equine with the

slowest being noted in avian strains (Bean, Schell et al., 1992). From rate comparisons, equine-1

A viruses are divided into two distinct groups from 1956-63 and 1964-77. The silent mutation

rate of HAl for the two periods was calculated to be 0.014%/year and 0.169%/year and the

amino acid substitution rate is 0.214%/year and 0.412%/year respectively (Gibson, Daniels et

al., 1992). This is consistent with other published data reporting that equine influenza viruses

isolated before 1968 had substitution rates of 3.1 nucleotides and 0.8 amino acids per year

whereas strains from 1968 and beyond accumulated 7.9 nucleotide and 3.4 amino acid

substitutions per year (Fields, 1996).

Neuraminidase. Neuraminidase (NA) membrane protein is a homotetramer that has a molecular

weight of 50,087, contains 1,413 nucleotides, and possesses both biological and antigenic

significance(Laver, Colman et al., 1984). The NA is a class II integral membrane protein which

consists of: 1) a trans-membrane stalk region with a high degree of variability between both A

and B virus and sub-types in regards to both length and amino acid sequence; 2) a head region

with varying degrees of homology among A and B and A subtypes; and 3) a cytoplasmic tail

which is conserved among A but not between A and B viruses (Fields, 1996). Each monomer is

made of 4-stranded anti-parallel beta-sheets which fold into six identical units. Each unit makes

up a binding pocket (Colman, Varghese et al., 1983). Neuraminidase is believed to be important

in initiating infection of cells of the respiratory system by facilitating transport of the virus

through the mucin layer. Its biochemical activity includes the ability to bind and remove sialic

acid by catalyzing the cleavage of alpha-ketosidic linkages between HA and NA and the host cell

surface (Palese, Tobita et al., 1974). Some NA cause hemagglutination; this feature has been

used to identify subtypes (Webster, Hinshaw, Bean, Van Wyke, Geraci, St Aubin, & Petursson,


M1 and M2. The viral matrix protein (MI) is the most abundant protein in the influenza virion

and is believed to be a structural protein underlying the lipid envelope. The Ml interacts with the

cytoplasmic tails of HA, NA, and M2 proteins as well as RNP structures (Rees & Dimmock,

1981). This was determined by a study that showed treatment with amantadine (1-

aminoadamantane hydrochloride) inhibited the flow of ions into the virion during uncoating,

thereby resulting in the dissociation of RNP from the Ml protein (Martin & Helenius, 1991).

M2 protein is a type II integral membrane protein homodimer with disulfide-linked pairs

that form a tetramer (Holsinger, Nichani et al., 1994). Several copies are synthesized during

replication with 20 to 60 molecules being integrated into the whole virus. These are expressed

during replication on the apical surface (site of virus budding) of the infected host cell. The M2

protein consists of N-terminal extracellular residues, a trans-membrane domain which is

responsible for forming the Na+ and hydronium permeable ion channel, and a cytoplasmic tail

(Pinto, Holsinger et al., 1992) The M2 molecule is an important target of anti-viral therapy with

amantadine (1-aminoadamantane hydrochloride) which disrupts ion transport (Oxford &

Galbraith, 1980).

Viral Transmission

The influenza virus particle makes multiple low-affinity attachments to sialic acid-

containing receptors through galactose, alpha 2,3 or 2,6 linkages (Weis, Brown et al., 1988). The

appropriate host cell, containing clathrin-coated sections of plasma membrane, engulfs the virus

particle through a process called receptor-mediated endocytosis, formerly referred to as viropexis.

Upon entry into the cell, the coated membrane-bound vesicle releases the virus by fusion into

endosomes. The endosomes provide an environment of decreasing pH necessary for the

processing of HA which leads to fusion of the viral envelope with the endosome membrane. In

some human strains, this process takes about 20 to 35 minutes and is required before the matrix

protein can release the RNP for transport into the nucleus via nuclear pores. However, before the

M1 protein can dissociate from the RNP, a requisite flow of ions from the endosome

compartment through the M2 ion channel into the virus core occurs. It is also postulated that the

flow of ions into the virus further "prepares" the transmembrane portion of the HA molecule for

fusion earlier on in this process. This is the point where antiviral drugs are targeted by utilizing

the fact that acidotropic weak bases such as amantadine and chloroquine will raise the pH within

endosomes (Matlin, Reggio et al., 1982). The idea of low pH leading to uncoating is consistent

with the more neutral pH at the cytoplasmic membrane, promoting virion assembly. Low pH is

proposed to facilitate the dissociation of the cytoplasmic tail of the HA protein from Mi protein

and intact RNP (Wharton, Belshe et al., 1994). This is followed by entry into the nucleus and

subsequent mRNA transcription of RNP.

RNA Replication and Protein Synthesis

The influenza virus is a single-stranded RNA virus equipped with some, but not all of

the proteins needed for replication. Within infected cells, viral RNA is transcribed into mRNA

and replicated. Synthesis of viral mRNA requires a host-cell RNA polymerase II as well as host

RNA as a source for primer sequences which are generated by viral endonucleases (Fields,

1996). A cap-dependent endonuclease will cleave sections from a methylated 5'-capped cell

RNA 10 to 13 nucleotides from the end, preferentially at a purine residue (Panniers & Henshaw,

1983). In the presence of viral RNA associated with NP, PGI, PB and PA, transcription can

proceed with a G residue incorporated at the 3' end (Inglis, Carroll et al., 1976). Specifically,

PB1 adjacent to the G residue is reported to catalyze the addition of individual nucleotides. PB2

is responsible for recognizing and binding to the cap primer sequence. Messenger RNA is

synthesized until the termination sequence of 5 to 7 uridine residues is reached (Rogers &

Paulson, 1983). A poly-A tail is then added to the 3' end of the nascent chain (Inglis, Carroll,

Lamb, & Mahy, 1976). Viral RNA replication not only includes the production of a template

strand, but its subsequent copying back into viral RNA.

It appears that primers derived from host-cell RNA polymerase II transcripts are required,

which explains the activity of actinomycin D and a-amanatin (Plotch, Bouloy et al., 1981). Viral

RNA replication is completed in two events: 1) the synthesis of full-length copies of an RNA

template; and 2) the subsequent copying of these templates into viral RNA. Influenza virus gene

expression is based on quantitative changes in the transcription of specific RNA segments (Lamb

& Choppin, 1976). Unlike DNA viruses, influenza virus gene expression occurs in distinct stages

that couples the synthesis of viral mRNA, viral RNA, and viral proteins (Hay, Abraham et al.,

1977). The selective copying of a particular template RNA into viral RNA results in the

subsequent production of mRNA and its corresponding protein. For example, NS and NP viral

RNA are synthesized early, whereas the Ml is produced at a later point during replication. These

events occur within the nucleus where the products remain until the appropriate time for egress

and assembly is reached. The timing of these events is closely associated with a sequence of viral

protein and RNA production. The "late phase" product, M protein, basically signals the end of

mRNA production and thus results in the assembly and transport of viral components from the

nuclease into the cytoplasm (Martin & Helenius, 1991).

Integral membrane proteins HA, NA, and M2 are synthesized in the cytoplasm on

membrane bound ribosomes. Trafficking of the proteins from the ribosomes through the trans

Golgi network (TGN) is via a signal recognition particle-dependent system. For HA proteins, a

transitional sequence of appropriate folding steps occurs as they progress through the ER and

Golgi (Boulay, Doms et al., 1988). At this point amantadine can cause a premature HA

conformational folding resulting in an extrusion of the fusion peptide at an inappropriate time

within the wrong sub-cellular compartment. This in turn inhibits the budding of an infective

influenza particle. Other post-transcriptional changes are made to the membrane proteins before

they exit the TGN. Complexed forms of HA and NA containing sialic acid and multi-basic

residues on their carbohydrate chains are cleaved in the TGN by an endoprotease, furin (Stieneke-

Grober, Vey, Angliker, Shaw, Thomas, Roberts, Klenk, & Garten, 1992). Later, integral

membrane proteins followed by Mi-associated RNP are transported to the apical-surface of the

cell membrane prior to budding (Hull, Gilmore et al., 1988). The release of a fully formed viral

particle requires neuraminidase; defective NA results in large aggregates of virus particles

(Palese, Tobita, Ueda, & Compans, 1974).

The Significance of Equine Influenza

The prevalence and incidence of equine respiratory disease has been well documented in

association with equine influenza virus (EIV). Understanding the pathogenicity of a disease

affecting a species that possesses such a significant recreational and economic value is of

paramount importance. Although there were suspect cases centuries ago, the earliest documented

case of an EIV outbreak in horses was in Prague, Czechoslovakia, in 1956; the virus isolated was

designated A/equine/Prague/1/56, (H7N7XSovinaova, Tumova, & Pouska, 1958). The next

significant subtype ofEIV to be isolated was A/equine/Miami/l/63 (H3N8) (Wadell, Teigland, &

Sigel, 1963). These two strains of EIV are commonly referred to as equine 1 and equine 2

respectively and the type 2 virus is considered to be more virulent (Wood, Mumford et al., 1983)

and possesses the potential to cause severe disease in young horses (Beech, 1991). While

antibodies persist in unvaccinated horses, no infections by the subtype A/equine/l have been

reported since 1979. However, a recent screening of 359 serum samples collected from

unvaccinated horses in Bosnia Herzegovina indicated that antibodies to A/equine/l/Prague/56

influenza virus were detected in 8.3% of the samples using the HI test (personal communication

to Paul Gibbs from Ramiz Velic).

Based on changes in the HA and NA genes, A/equine/2 strains show a division into two

subgroups. The newly characterized H3N8 virus underwent sufficient antigenic drift at the HA

and NA level to result in two variant subtypes, one including the A/Miami/63 (prototype),

A/Tokyo/71, and A/Switzerland/79 and a second represented by A/Fontainblue/79,

A.Newmarket/79, A/Solvalla/79, and A/Kentucky/81 (prototype variantXWilson, 1993).

Increased efforts in surveillance for the A/equine/2 virus has resulted in the recognition of

geographic divergence into two further categories, the American-like (isolated from horses in

USA and Argentina) and the Eurasian-like (isolated from horses in Europe, Mongolia and China)

(Wilson, 1993).

Host Specificity

While there are examples of "host specific" influenza viruses in mammals such as those

seen in humans, whales, pigs, and horses, they are believed to be related to strains of avian virus

(Hinshaw, Bean et al., 1986;Webster, Hinshaw, Bean, Van Wyke, Geraci, St Aubin, & Petursson,

1981;Guo, Wang et al., 1992). Evidence for a conserved lineage between swine and human

influenza strains is found in the fact that both have similar NP genes arising from a common

origin. Additionally, NP genes of an avian origin have been determined in mink,

A/Mink/Sweden/84, swine A/Swine/Netherlands and A/Swine/Germany/81, and horses

A/Equine/Jilin/89 (Fields, 1996). One study showed that the amino acid sequences of the HAl

molecules of A/duck/Ukraine/63 and A/equine/Miami/63 had similar "silent-mutation" rates.

From this, Daniels et al. deduced that a common ancestral virus existed around 1945 (Daniels et

al., 1985).

The pathogenicity of a particular strain is closely related to the HA protein. The

attachment of virus to the host-cell is accomplished through interactions between the HA and

sialic acid-containing glycoproteins, specific for an alpha 2:3 linkage. Furthermore, host

specificity is a result ofa glutamine residue in animal strains (seen in A/equine/Miami/63,

A/equine/Fontaineblue/79, and A/duck/Ukraine/63) instead of a leucine seen in the A/ human X-

31 virus/69 at site 226. This is an example of a point mutation that can have relevance to

antigenic drift as discussed later.

Another possible explanation for the heterogeneity of the various strains of influenza is

through evolutionary pressure exerted by the immune system of the host species, sometimes

referred to as host immune-selection pressure." This occurs most dramatically where the

number of replications is high. Because humans provide a large reservoir of susceptible hosts,

the number of replications is high, allowing the HA surface protein to evolve at a much quicker

rate than other species (Bean, Schell, Katz, Kawaoka, Naeve, Gorman, & Webster, 1992). In

comparison, nucleotide changes in influenza of avian origin occur much more slowly and do not

result in amino acid changes. This silent mutation in avian viruses has resulted in a phylogenetic

profile that apparently has not changed in over 50 years (Gorman, Bean et al., 1991). This

concurs with the idea that conservation within the avian species is a result of these viruses

reaching their adaptive potential where nucleotide changes afford no advantage for selectivity.

This has significance relative to human infection; some of the viral gene segments that caused

pandemics in humans in 1918, 1956, and 1968 still persist within avian species with few or no

mutations (Gorman, Bean, Kawaoka, Donatelli, Guo, & Webster, 1991).

Shift and Drift

Antigenic shift is the result of reassortment of genes between viruses of two strains.

Reassortment occurs in viruses with segmented genomes and has a low rate of occurrence. It can

result in the emergence of a highly virulent strain (Rott, Orlich et al., 1984). An example of this

was the reassortment of genes between avian and human virus strains to produce the 1957 Asian

"flu" outbreak. This new, antigenically distinct strain derived its HA, NA, and PBI genes from

an avian virus while the remainder of the genes were of human origin. (Kawaoka & Webster,

1989;Scholtissek, Koennecke et al., 1978). Other reassortments have occurred between other

mammals and avian species (Guo, Wang, Kawaoka, Gorman, Ito, Saito, & Webster, 1992).

Within a period of little more than a year, two new strains of EIV were characterized from the

northeastern region of China (Guo, Wang, Kawaoka, German, Ito, Saito, & Webster, 1992;Guo,

Wang et al., 1995). Emerging strains of EIV, as in human influenza, can cause severe clinical

disease based on the absence of serum antibodies that recognize antigenically similar strains. The

two strains isolated in northeastern China were determined to be antigenically similar to each

other and to a H3N8 subtype A/equine/2/Miamill/63 strain and were referred to as Jilin/l/89 and

Heilongjaiang/15/90. Phylogenetic analysis of complete and partial sequences of the HA, NP, M,

and NS genes revealed that Jilin/1/89 was most similar to an avian H3 virus,

A/Duck/Ukraine/3/63. While the reassortment of genes can cause pandemic viruses to emerge,

another possibility is that a virus that has been out of circulation for a number of years can

emerge again due to the birth of a new susceptible population in the intervening period. This was

speculated as being the cause of the 1977 pandemic of human influenza in China due to a Russian

strain which had not been seen for 27 years (Nakajima, Desselberger et al., 1978).

RNA viruses, which include EIV, have the capacity to undergo spontaneous mutations

(point mutations) which result in phenotypic changes (Fields, 1996). This is called antigenic

drift. In one study, 79% of changes in the amino acid sequence between two equine strains

(Miami/63 and Fontainebleau/79) were in the HAl gene, which highlights the importance of the

antigenicity of HAl (Daniels, Skehel, & Wiley, 1985). The rate of changes that occurs is highest

in humans, followed by equine and swine, with the lowest being found in the highly conserved

avian strains. In a study conducted by Skehel and others, the rate of amino acid substitutions in

avian species was 0.36%/oyear over a 16-yr period as compared to human influenza which is about

0.87%/year over 14 yr (Skehel, Daniels et al., 1983). Silent mutations consisting of nucleotide

changes that do not result in amino acid substitutions in the HAl coding region, follow a similar

trend (0.17% for horses and 0.3% for humans). The rate of changes is presumed to be a function

of the number of replication cycles and not due to species-specific factors. There are five

antigenic sites on the HAl protein A to E which are most susceptible to changes resulting from

point mutations. The nucleotide changes in these distal regions of the molecule HAl are the basis

for virus evasion from antibodies generated in the host by previous exposures antigenicc drift).

An interesting point is that immunologically induced pressure is believed to affect the selection of

mutants in human and other mammalian species whereas in avian species, this is less the case

(Air, Gibbs et al., 1990). This is due to the fact that avian viruses have an evolution rate

(nucleotide change rate) that does not result in amino acid changes. Because the nucleotide rate

is slower and does not lead to amino acid changes, it appears that avian influenza virus has

reached its optimum capacity to adapt.

Mode of Transmission and Infection

The natural route of infection for horses is through direct transmission by aerosol or

indirectly through fomites (Ames, 1988). Aerosol transmission is the more important route.

Experimental infection protocols have included intranasal inoculation and aerosolized mist with

nebulizers as a means of administering live influenza virus (Mumford, Wood et al.,

1983;Mumford, 1991;Mumford, Wilson, Hannant, & Jessett, 1994). In humans, natural and

experimental infections are through person-to-person contact or via the aerosol route respectively

(Douglas, 1975). The number of infectious particles required to produce clinical signs in horses

has been investigated by Mumford and others with experimental doses ranging from 106

EID50/mL using a nebulizer and 1087 EIDso/mL dose during intranasal instillation (Mumford,

Wood, Scott, Folkers, & Schild, 1983;Hannant, Jessett et al., 1989). During natural infection, it is

speculated that fewer virus particles are needed to disseminate the disease. It is rare to isolate a

titer of more than 103 EIDso/mL from horses naturally infected with EIV (Mumford, Wilson,

Hannant, & Jessett, 1994). The average size of a particle generated by humans during a sneeze or

cough is less than 2 micrometers in diameter and could contain 1 TCIDso, most likely deposited in

the lower respiratory airways of the lung (Knight, 1973). The explosive cough from a horse with

EIV can result in the effective aerosolization of viral particles. No studies have been conducted

to determine the size of aerosolized particles generated from horses or the amount of virus

contained in each. However, what has been well established is that the target cells of EIV are in

the upper and potentially lower respiratory tract (Beech, 1991).

Clinical Presentation

Under optimum conditions of transmission, where infected (possibly sub-clinically)

horses co-mingle with those that are susceptible, the onset of influenza can be acute and the

dissemination rapid (Ames, 1988). The virus is shed during the incubation period and horses can

remain infectious for up to 5 days after the appearance of clinical signs (Murphy, Gibbs,

Horzinek, & Studdert, 1999). Elevated temperatures from 39.1 to 43.3 0 C can last from 2 to 5

days accompanied by a serous-type nasal discharge. Often this precedes a secondary bacterial

infection with an associated mucopurulent discharge (Ames, 1988). In some cases, mortality can

be a result of bacterial as well as viral pneumonia (Baker, 1986). In a 1994 epidemic in China,

30,000 horses died and 1.5 million became seriously ill when bacterial infections complicated an

EIV outbreak. (Shortridge & Watkins, 1995). An unproductive, yet explosive, cough is harsh

and can persist for up to 3 weeks. Anorexia, depression, tachypnea, tachycardia, and myalgia

can also be associated with clinical disease related to EIV (Coggin, 1979). Reports of

cardiomyopathy have been noted in rare cases. In humans, myositis with an increase in serum

muscle enzymes (myoglobinemia and myoglobinuria) and accompanied by pain and swelling of

the limbs can occur (Middleton, Alexander et al., 1970). Similar signs have been documented in

horses; they appear stiff and are hesitant to move with a slight swelling in the limbs (Beech,

1991). During a recent study, horses which where seronegative to EIV (A/Eq/Prague/56,

A/Eq/Miami/63, A/Eq/Saskatoon/90, and A/Eq/Kentucky/91) were challenged with an aerosol of

A/Eq/Kentucky/91 at 1068 EIDso/animal (Morley, Gross et al., 1998). After experimental-

infection, some horses were exposed to training on a treadmill and some were rested in their

stalls. Horses exercised during the influenza infection showed an increase in the severity of

clinical signs. Specifically, fever, coughing, mucopurulent nasal discharge, anorexia, and

lethargy persisted longer and were more severe in the exercised horses infected with EIV than

those which were rested. Additionally, exercised horses lost more body weight than did their

cohorts. Another study showed that permanent damage to the myocardia can occur through

interstitial myocarditis present in horses after infection with A/Equine/Miami/l/63 (Gerber,



The pathogenesis of influenza has been described for humans (Small, 1990) and horses,

(Beech, 1991;Coggin, 1979). Virus particles pass through the mucous layer and adhere to ciliated

columnar epithelial cells. These cells appear vacuolated, edematous, and lose cilia before they

are desquamated (Beech, 1991). Damage to the respiratory epithelium results in a decrease of

respiratory clearance leading to secondary bacterial infection in humans (Small, 1990). Similar to

humans, horses are susceptible to secondary respiratory-tract infections (Beech, 1991). Around 1

day post-infection, the number of mucous producing and ciliated cells will be reduced to the point

where areas of thickened and hyalinized basement membrane are exposed. Sub-mucosal edema

and hyperemia results from infiltrating inflammatory cells, neutrophils and mononuclear cells

(Coggin, 1979). Additionally, alveolar macrophage function is reduced (Ames, 1988). At 3 to 5

days after the onset of illness in humans, regeneration of the epithelium is initiated by the

differentiation of basal cells into serious and ciliated cells (Small, 1990). What is both interesting

and biologically practical is the fact that basal cells are not susceptible to infection by influenza

virus. If this were indeed the case, infection in humans and animals with the influenza virus

would be fatal to the host and self-limiting. The apparent reason for the inability of basal cells to

become infected is due to their lack of the proper receptors for influenza (Small, 1990).


Outbreaks can involve horses of varying ages and immunity to EIV but are most

commonly seen among animals from I to 3 years of age (Ames, 1988;Livesay, O'Neill et al.,

1993). An outbreak of upper respiratory disease (URD) in standardbred horses in Canada over

the 2-yr period of 1973 to 1974 showed an age-specific infection associated with EIV. In the first

and second years of the outbreak, 2 yr-old horses had the highest attack rate among a group of 2

to 9 year-olds at 0.24 and 0.25 for each year respectively (Sherman, Mitchell et al., 1979). As

would be expected, the severity of the infection was reported to be dependent upon prior

immunity to EIV. Horses with pre-existing serum antibody titers to EIV will have less severe

clinical signs and will shed less virus for a shorter period of time (Hannant, Mumford et al.,

1988). It seems reasonable that outbreaks have a better chance to occur when sub-clinically

infected horses are mixed with naive young horses not previously exposed to or vaccinated

against EIV. This situation often presents itself at race-tracks, shows, and breeding farms. In the

previous example, Sherman et al., 1979 described meets held in Ontario, Canada during winter,

spring, and fall where standardbred horses had episodes of URD associated with equine influenza

and rhinopneumonitis virus. As reported with the human strains of influenza (Fields, 1996), EIV

survives in aerosol in a relatively low humidity, low temperature environment which can be seen

in stables at holding facilities or during the winter months (Beech, 1991).

Despite the existence of vaccines over the last 30 years (Bryans, Doll et al., 1966), EIV

has circulated within equine populations in North America, Europe, and Scandinavian (Powell,

Watkins et al., 1999). Recent epidemics (i.e. last 20 years) have been reported in the United

Kingdom, China, India (Mumford, 1999a), Hong Kong (Powell, Watkins et al., 1995) and Croatia

in 1993 (Madic, Martinovic et al., 1996). When horses are mixed at racing, training, or breeding

facilities, their origin of travel can impact the potential for an outbreak. During an outbreak at the

Hong Kong Jockey Club in 1992, individual or groups of animals were identified that had

recently arrived from the UK and Ireland, countries where horses encounter EIV and,

accordingly, are routinely vaccinated. Horses from Australia and New Zealand, where EIV is not

prevalent, were stabled at the same facility. These horses had not been vaccinated against EIV

until their arrival at the race-track facility. The incidence of disease was much higher (52%)

among the horses from Australia and New Zealand as compared to those from the UK and Ireland

at (20%). This outbreak represents an example of the conditions where equine influenza can

spread among a susceptible population of horses. Thus, it appears that the origin and destination

of horses were key to understanding the risk factors associated with the 1994 epidemic at the

Honk Kong Jockey Club.

Impact to the Industry

Horses infected with the EIV typically are symptomatic but recover with good

management. Why then is there such concern over this disease in the equine community? If

the horse population is categorized into those that are "recreational animals" and those that are

"performance animals", one can better appreciate the impact of influenza. It is the economic loss

associated with influenza outbreaks in performance animals that makes the disease important.

With the expansion of international trade in horses and equine products (semen), comes the

associated economic significance, mirrored by the potential spread of infectious agents. With the

exception of humans, horses travel internationally (usually by air) more than any other animal

species (Timoney, 2000). Because of the potential risk of reduced or lost performance of racing,

jumping, evening, show, or breeding equids, much emphasis has been placed on monitoring and

controlling infectious upper respiratory disease (IURD) in horses. Thus, to appreciate the

impact of influenza on the equine community, both performance and economics can be


Horses that are sub-clinically infected with EIV not only cause endemic disease, but also

are at risk of training or performance failure. Movement of sub-clinically infected horses at the

international level provides the potential for interaction with susceptible populations (Timoney,

2000). As with other infectious diseases, vaccination strategies are always of primary

importance. Current methods for vaccination, while moderately effective at controlling clinical

illness and the spread of infection, do not provide a tool for complete eradication (Mumford,

1999a). This is due to not only the type of vaccines routinely used, but also related to the

continuous antigenic changes that occur with influenza virus. Because of this, individual events

of URD associated with EIV are frequent and result in the loss of performance. The second most

common cause of loss of training days for racehorses is a result of URD from various etiologies

(Wood, Newton et al., 1999). Outbreaks of IURD resulting from infection with equine herpes

virus-1 and 4 (EHV) and EIV can occur when horses are congregated at places such as a

"pinhooking" facility. The spread of IURD among yearlings can cause a significant decrease in

performance associated with training failure. In a recent survey, three out of 76 thoroughbreds

(4%) followed at a pinhooking facility in Ocala Florida were diagnosed with clinical signs of

IURD (Hernandez, J. personal communication).

A second impact of EIV is the loss of revenue associated with an outbreak. In 1975, the

cost resulting from infection with EIV was estimated to be $668 per horse (Bryans, 1975).

Watkins reported the loss of revenue in a more recent outbreak during the 1992 racing season at

the Hong Kong Jockey Club. The cancellation of seven race meetings resulted in the estimated

loss of nearly $ Ibillion US dollars, of which, 15% would be collected in the form of tax revenue

to the local community (Watkins, 1997). To further appreciate the economics associated with

horse racing, Table 1-2. illustrates an example of the 1995/96 racing season in Hong Kong.

TABLE 1-2. Economic losses associated with an outbreak of EIV at the Hong Kong Jockey Club
during the 1995/1996 racing season

* 69 race-meetings
* 517 races
* 6,090 starters from an average population of 1,000 horses
* US $ 92 billion turnover for the season
* US $ 130 million average turnover every race meeting
* US $ 17.5 million average turnover every race
* US $ 1.5 million bet on average on every horse that raced
* US $ 1.3 billion returned to the community in tax, duty and donations
* Total attendance of 3,270,100 with an average of 47,400 every racemeeting
* Total employees of 16,178

Taken from Watkins, 1997

Vaccination Against Equine Influenza Infection

Vaccine Theory

Vaccines have been developed in an attempt to eliminate or control the spread of

infectious agents as well as to moderate clinical disease. Vaccination not only eliminates some

discomfort to the animal, but has been shown to be a cost-effective way to minimize economic

losses as well. Vaccination, at the most basic description, is the delivery of an immunogenic

antigen (live, killed, or parts of a pathogenic organism) to the host resulting in a protective

immune response. The ideal vaccine should be efficacious, economic to manufacture and

administer, and have minimal adverse side effects. A great deal of progress in the fields of

genetics, virus structure, and the immune system has allowed the development of effective

vaccines. New approaches to vaccines used in veterinary medicine must result in those that

minimize the undesirable side-effects, are efficacious, and safe to the vaccinate well as non-

vaccinated animals and humans.

There has been a great deal of progress in the standardization of vaccine testing relative

to EIV. Vaccines against EIV have a direct correlate between HA content and the level of

protection. Because of this, the hemagglutination inhibition assay has been a gold standard to

determine the efficacy of vaccines (Hoskins, 1967). Another reliable in-vitro measure of serum

antibody is the single radial hemolysis (SRH) assay (Wood, Schild et al., 1983). Both have been

used extensively during the development of vaccines against EIV.

Early versions of vaccines against EIV included inactivated virus vaccines with various

adjuvants. Inactivated vaccines are still in wide use today. However, inactivated-virus vaccines

provided incomplete and short-lived protection against severe clinical disease. Furthermore, they

are unable to prevent infection. While improvements have been made with adjuvants that result

in a prolonged antibody response, alternative methods to present antigen in a manner that more

closely resembles natural infection are needed. In an effort to mimic complete protection

provided by natural infection, novel approaches in vaccine development have included modified-

live virus (MLV) and recombinant DNA vaccines. Presently, vaccine development, along with

international efforts to improve surveillance are identified as priority issues for the control of

EIV (Mumford, 1999a).

Killed Vaccines

Currently, inactivated-virus influenza vaccines are licensed for parenteral administration

in humans (Fields, 1996) and animals (Horzineck, Schijns.V.E. et al., 1997). A killed or

inactivated vaccine constitutes an infectious agent which has been altered by chemical or other

methods, such that replication is not possible, yet antigenically important proteins are conserved.

Whole virus and subvirion vaccines contain intact, inactivated virus, or purified virus disrupted

with detergents that solublize the lipid-containing envelope. Unfortunately, some inactivating

compounds (formalin or beta-propriolactone) can alter the immunogenicity of the antigenic

proteins (Duque, Marshall et al., 1989).

One of the greatest shortcomings of inactivated vaccines is an incomplete immune

response. While inactivated-virus vaccines can elicit a good humoral response, they are not

effective at inducing cell-mediated or mucosal immunity. However, the use of various lipid

containing structures or immune stimulating complexes (ISCOMs) may produce both a local and

humoral response when used with inactivated vaccines (Rimmelzwaan & Osterhaus, 1995).

Immune-mediated resistance induced by experimental or natural infection with influenza

virus is based on the level of protective HA antibodies. The same level of protection is not

achieved with the use of inactivated-virus vaccines. This is of particular importance in humans

and horses where HA surface antigens contained in a vaccine may not be effective against the

strain circulating in an epidemic (Mostow, Schoenbaum et al., 1970).

Live Vaccines

Live vaccines have several distinct advantages over those that are inactivated. They

possess properties that allow various routes of inoculation, they present a complete suite of viral

antigens, they can induce cell mediated and humoral responses, and they are economical to

produce relative to the cost of inactivated-virus vaccines. Live vaccine are not without

drawbacks, however. The associated disadvantages include immunosuppression and the potential

to cause and spread disease among immunocompromised individuals. Modified live virus

vaccines have been the subject of experimentation in humans for use as an immunoprophylaxis

against influenza for several years (Advisory Committee on Immunization Practices (ACIP),


The exact mechanism by which a virus becomes attenuated during serial passage is not

known. However, the genes responsible for the production of toxins, structural proteins, and

nucleic acid metabolism can be altered, and result in attenuation (Muster, Subbarao et al., 1991).

One of the potential problems with attenuated viruses is their potential to revert back to a virulent

form. In some vaccines, attenuation is the result of point mutations that can revert back upon a

single passage in vivo (Minor, John et al., 1986). An example of this occurred during a study

that used the H3N8, A/equine/Miami/63 virus at the 6e passage as a challenge-strain in horses.

The virus was passage five times and inoculated into two sero-negative ponies. Although virus

replication occurred, neither pony showed clinical signs. The virus from one of the ponies was

inoculated into the nasopharynx of a third seronegative pony and clinical disease ensued

accompanied by fever and coughing. The same virus strain was again isolated from the horse and

passage once in eggs. A subsequent challenge using the egg isolate at 10S EIDs0/mL was

capable of inducing clinical signs (Mumford, Wood et al., 1988).

New approaches in the development of live vaccines for influenza must take antigenic

drift into account. For this reason, it is not practical to produce an attenuated form of each

subtype that emerges during an outbreak. To address this problem, site-directed mutagenesis is

used to produce an attenuated virus with changes in genes such as NA (Muster, Subbarao, Enami,

Murphy, & Palese, 1991). In one study, a wild-type virus was passage using a method that

resulted in a mutation. The attenuated virus was mixed with a wild-type virus and the reassortant

progeny proved to be effective and safe (Edwards, Dupont et al., 1994).

Recombinant DNA Vectors

A recombinant DNA vaccine will typically consist of an avirulent organism that is

constructed by site-directed mutagenesis to carry a gene insertion encoding for an immunogenic

antigen. This approach also assumes the role of the host to replicate the vector organism (Perkus,

Piccini et al., 1985). A safe and effective virus vector will possess a large genome consisting of

non-essential genes that can be deleted. Furthermore, the gene-deletions should result in a non-

virulent virus that retains the ability to replicate. As knowledge about gene insertions and

deletions grows, so will the number of vaccines that carry this technology. Large DNA viruses

such as vaccinia virus are suited for this construction (Cooney, Collier et al., 1991). One study

included a vaccinia virus and a plasmid vector used to express surface glycoproteins encoded by

two strains of EIV types 1 and 2 (Beverly, Brown et al., 1988). In this study, a recombinant

vaccinia EIV vaccine was used that expressed HA and NA. They found that an initial serum

antibody response rapidly declined in manner similar to that seen in naturally infected horses.

They also concluded that, in addition to immunoglobulin production, cytotoxic T lymphocytes

(CTL) are partially responsible for the protection provided by a recombinant vaccinia virus

vaccine. In similar studies, hamsters and mice were reported to have an increased CTL response

following inoculation with a recombinant vaccine expressing A/Japan/305/57 antigen (Smith,

Mackett et al., 1983).

Immune Response

The impetus for generating an antigen-specific response that provides protection against

the harmful effects of a pathogen (vaccination) assumes the basic requirement of a functioning

immune system. The induction of a protective response is referred to as acquired immunity. A

second form of protection comes from innate or natural immunity which describes non-immune

barriers such as respiratory mucous, mucociliary transport and the cough reflex (Dixon &

McGorum, 1997). Vaccines are effective at generating acquired immunity against a pathogen.

The type of immune response is dependent upon the preparation of the antigen and can be either

cell-mediated or systemic or a combination of both.

Antigen Presentation

More specifically, the nature of the immune response is dependent upon the way antigen

is processed and presented by antigen presenting cells (APC) for recognition by specialized

effectors cells. This is the basis for the increased efficacy of whole-virus preparations over that

of inactivated or subunit vaccines. Studies have shown that vaccines containing MLV induce

several appropriate effector mechanisms that lead to a comprehensive response (Lunn, 1997).

Antigen in its native form is not recognized and acted upon by all immune cells. B cells

recognize native antigen through a receptor associated with immunoglobulins on their surface. In

contrast, T cells responsible for functions related to cytotoxicity and antibody production, require

antigen to be presented by specialized APC which process antigen into recognizable peptide

segments. Examples of APC are monocytes, macrophages, B cells, and dendritic cells (Dixon &

McGorum, 1997). Before T-cell recognition can occur, antigen peptides must be associated with

major histocompatibility complex (MHC) molecules through either an endogenous (MHC I) or

exogenous (MHC H) pathway. The subtypes of T cells that respond to antigen include CD4' T

helper (Th) via the MHC class I exogenous pathway or CD8+ CTL through the MHC class I

endogenous pathway (Sprent & Tough, 1994;Accolla, Auffray et al., 1991). Virus-infected cells

will process viral-peptide fragments that are expressed on the surface in association with MHC

class I class molecules resulting in a Thl-like response (Monaco, 1992). This results in the

differentiation and activation of CTL effector cells and T helper lymphocytes that, in turn,

stimulate antibody-producing cells (Dixon & McGorum, 1997). Typically, influenza antigen that

is presented in an inactivated-virus vaccine does not enter the endogenous pathway and is unable

to activate CD8+ T cells. Instead, viral peptides are presented on the cell surface in association

with MHC class II molecules resulting in, primarily, a Th2 response and antibody production

(Rimmelzwaan & Osterhaus, 1997).

Duration of Antibody Post-infection or Vaccination

Vaccines are based on the presentation of parts or whole pathogenic organisms to induce

an immune response. Protection against severe clinical disease during infection with influenza is

based on the formation of antibodies to HA and NA surface glycoproteins. Cell-mediated

immunity is typically directed toward the recognition of the NP of the influenza virus and is also

believed to play an important role in recovery after infection. Despite the importance of both

humoral and cell-mediated immunity, the primary surrogate of protection is the measure of short

and long-term circulating immunoglobulin proteins. IgG and IgM can provide protection by

neutralization of the virus through antigen-specific binding, opsinization to enhance the removal

of antigen by phagocytes, and the activation of complement (Dixon & McGorum, 1997).

Historically, the standardization of protocols to determine the efficacy of influenza

vaccines has been limited by the variability in assay methods. One study compared the ability of

the HI and SRH assay to detect a dose-dependant response of a vaccine against EIV (Wood,

Schild, Folkers, Mumford, & Newman, 1983). This study confirmed the greater specificity of

SRH over that of the HI test. Antibody titers were low and undetected by HI after a primary

inoculation with a bivalent vaccine containing inactivated A/equine/Prague/56 and

A/equine/Miami/63. However, low levels of SRH antibody were detected after the first

vaccination and a significant rise was detected after a booster inoculation. Another study

correlated HA units/dose and antibody titers associated with protection against severe disease

(Mumford, Wood, Folkers, & Schild, 1988). The results were confirmed with a SRH test which

showed a correlation between protection against clinical disease and a pre-challenge minimum

zone of hemolysis equaling 74 mm2.

In a more recent study, ISCOMS were found to be effective as an adjuvant during

vaccination against EIV challenge (Mumford, Jessett, Dunleavy, Wood, Hannant, Sundquist, &

Cook, 1994a). Two vaccines, consisting of HA antigen from A/equine/2/Newmarket/77 or

A/equine/2/Brentwood/79, were completed with an ISCOM adjuvant (tetanus toxoid adsorbed on

aluminium phosphate). A dose-related response established that a minimum of 15 micrograms of

HA/dose would provide a protective and slower declining serum antibody titer. Antibody titers

associated with protection were not achieved until after the third inoculation of vaccine, but were

maintained for up to 15 months. The use of the ISCOMs increased antibody-mediated protection

against clinical disease. However, while local immunity and cell-mediated immunity were not

investigated in this study, they may have contributed to the resulting increased protection. The

use of ISCOM adjuvants has been shown to facilitate the presentation of antigen in association

with MHC class I molecules to T-cell receptor (TCR) molecules on CTL. The use of ISCOMs

will result in local antibody production which is most likely the most important mechanism of

prevention (Horzineck, Schijns.V.E., & Denis, 1997). As stated previously, even though

antibody production is considered key to protection, several reviews highlight the importance of

the local and cell-mediated response during EIV infection (Lunn, 1997; Hannant and Mumford,

1997; Plateau et al., 1997; Dale et al., 1997; Mumford et al., 1983; Mumford et al., 1994).

Mucosal Immunity

The mucosa of the respiratory tract is lined with tissue responsible for protection against

foreign antigens as the first line of defense (Small, 1990). Mucosal immunity is important in

preventing infection through local antibody and cell-mediated responses (Bender & Small, Jr.,

1992). In the mucosal lining, antigen is taken up by specialized epithelial (M) cells which

mediate its transport to mucosal-surface-associated lymphoid tissues (MALT) (Horzineck,

Schijns.V.E., & Denis, 1997;McGhee, Mestecky et al., 1992). T-lymphocyte helper cells are

thought to be the source of localized cytokines IL-5 and IL-6 involved in class switching

(Horzineck, Schijns.V.E., & Denis, 1997;McGhee, Mestecky, Dertzbaugh, Eldridge, Hirasawa, &

Kiyono, 1992). Within the MALT, antigen presentation will result in class switching of

immunoglobulins and the activation ofIgA-specific memory B cells. IgA secreted at the mucosal

surface will effectively bind foreign antigen to prevent attachment to the epithelial cell surface

(Holmgren, Czerkinsky et al., 1994). In addition to the neutralization of virus in the lumen by

IgA, the transcytosis system has two additional effector mechanisms. IgA can effectively

neutralize viruses within the epithelial cell as well as the lumenal surface (Mazanec, Coudret et

al., 1995a). Intracellular neutralization of virus has been demonstrated experimentally by

infection ofMadin Darby canine kidney cells (MDCK) with influenza or Sendai virus. There is

evidence that IgA can inhibit viral assembly (including influenza) within the epithelial cell by

binding viral proteins released from the Golgi (Mazanec, Nedrud et al., 1993). Newly

synthesized viral glycoproteins HA and NA move from the trans-golgi network to the lumenal

surface via endosomal compartments containing polymeric IgA (pIgA). Yet another mechanism

of protection mediated by IgA molecules is the inter-cellular transport of antigen across the

luminal surface of epithelial cells coupled by a secretary component (Mestecky & McGhee,


Again, an important component of vaccine-strategy is to induce a comprehensive immune

response, including the production of local IgA. In mice (Tamura, Funato et al., 1990;Tamura,

Funato et al., 1991) and humans (Boyce, Gruber et al., 2000;Tomoda, Morita et al., 1995), the

production of nasal IgA antibody has been associated with the use of MLV vaccines and

infection. In contrast, intramuscular administration of inactivated virus preparations do not result

in nasal IgA antibody production (Clements, Betts et al., 1986). In the horse, local IgA antibody

production was not detected after immunization with an MLV vaccine (Lunn, Hussey, Sebing,

Rushlow, Radecki, Whitaker-Dowling, Youngner, Chambers, Holland, Jr., & Horohov, 2001).

Furthermore, similar to humans and mice, the use of an inactivated-virus vaccine in horses does

not induce nasal IgA formation (Nelson, Schram, McGregor, Sheoran, Olsen, & Lunn, 1998).

Mucosal inoculation with a DNA vaccine, containing A/equine/2 proteins, resulted in the

production of nasal IgA antibody in ponies (Lunn, Soboll, Schram, Quass, McGregor, Drape,

Macklin, McCabe, Swain, & Olsen, 1999).

Hypothesis and Objectives


The intranasal or intramuscular inoculation of a novel recombinant DNA vaccine

constructed of an EHV-4 vector encoding HA and NA genes from A/equine/2/Kentucky/94 for

antigen presentation will induce systemic and local antibody production as well as the cell-

mediated arm of the immune response. The recombinant vaccine will provide protection against

infection in the form of reduced virus shedding and clinical disease.

The goal of this study is to show that the use of a novel recombinant DNA vaccine will

provide a similar immune response compared to that seen after natural infection or immunization

with a cold-adapted modified-live virus vaccine. Furthermore, immunization with the

recombinant vaccine will result in a greater decrease in viral shedding and clinical signs over that

seen after immunization with an inactivated-virus vaccine upon challenge-infection.


The objectives of this study are: 1) to further develop existing techniques to sample and

assay horses for the presence of serum and nasal antibodies and cell-mediated immunity to EIV,

2) conduct animal studies to determine the immune response of horses after immunization with

either a recombinant DNA, inactivated-virus, or modified-live virus (MLV) vaccine, 3) determine

the ability of each vaccine to eliminate or reduce virus shedding and clinical disease, and 4)

compare the vaccine-induced response to that seen in horses that are experimentally infected with





Studies related the development and testing of vaccines against influenza have

incorporated a number of methods to sample and assay for surrogates of protection. These have

included serology, nasal antibody, cell-mediated immunity, virus shedding, and clinical diagnosis

(Asanuma, Aizawa et al., 1998;Bender, Johnson et al., 1991;Hannant, Easeman, & Mumford,

1999;Hannant & Mumford, 1989;Lunn, Soboll, Schram, Quass, McGregor, Drape, Macklin,

McCabe, Swain, & Olsen, 1999;Mumford, Wood, Folkers, & Schild, 1988). Previous studies

have shown a correlation between the formation of antibodies and protection against severe

clinical disease in horses. Serum antibodies have been recognized as a surrogate marker of

protection and are the basis for measuring vaccine efficacy against infection with EIV.

Therefore, various assays were used to confirm the response of horses to vaccination and

infection. Hemagglutination inhibition, SRH, and neutralization assays are regarded as the "gold-

standards" for measuring serum antibody in response to influenza infection. These assays were

used as reference assays during the standardization of an ELISA. Optimization of the ELISA

will include virus growth and purification, determination of sensitivity and specificity, and further

development of existing techniques to sample and assay for nasal antibodies. Sensitivity and

specificity of an assay relates to the probability of testing positive or negative while that is, in-

deed, the true condition. In the present study, sensitivity and specificity of the ELISA were used

to determine a "critical" or cutoff value associated with seroconversion. Here, a "Critical-Value"

plot, similar to those used in Receiver-Operator Characteristics (ROC) plot analysis was used to

graphically display the sensitivity and specificity of the ELISA at various critical or cutoff


Several reports have included CMI analysis to characterize its importance in mice

(Bender, Johnson, & Small, 1991;Bennink, Yewdell et al., 1984;Epstein, Lo et al., 1998), humans

(Bernstein, Gardner et al., 1998;Kruse, Moriabadi et al., 2001), cats (Song, Collisson et al., 1992),

cattle (Abdy, Howerth et al., 1999), and horses (Ellis, Bogdan et al., 1995) (Ellis, Steeves et al.,

1997;Hammond, Cook et al., 1997)during their recovery from viral infections. Cell-mediated

immune function tests include lymphocyte proliferation, cytotoxic T-lymphocyte assays, and

cytokine determination. The lymphocyte proliferation assay has been routinely used in human

immunology as an indicator of basic immunocompetence (Fletcher, Klimas et al., 1992). The

lymphocyte proliferation assay is a good indicator of immune function in general, or to a specific

pathogen, but often interpretation of results is difficult due to variability in the data. As with

humans, this is the case with horses (Ellis, Bogdan, & Kanara, 1995). The methods used to

collect, cryopreserve, culture, and stimulate lymphocytes can vary. Recent studies characterizing

the blastogenesis of cells after stimulation have included the use of a tetrazolium dye (Owen's

reagent) (Behl, Davis et al., 1994;Lappalainen, Jaaskelainen et al., 1994). This method has been

compared to and used in place of a [3H] thymidine uptake assay (Wong & Goeddel, 1994). In

that study, the tetrazolium dye method demonstrated comparable sensitivity to the [3H]

thymidine uptake assay. Due to the advantages of using a non-radioactive reagent, the ability of

the tetrazolium dye to detect increases in equine PBMC numbers during a proliferation assay was


The role of cytokines in the regulation of immune function during viral infection has been

described elsewhere (see review) (Guidotti & Chisari, 2000). Limited information is available on

the characterization and generation of monoclonal antibodies to equine cytokines (Horohov,

1999;Lunn, Sobol et al., 1999) and ELISA techniques to measure cytokine production in horses

have not been described. However, studies are ongoing that describe the development and future

application of these reagents (Lunn, Sobol, Swiderski, Horohov, & Olsen, 1999). Generation of

primers to detect equine cytokine mRNA expression in a PCR assay has been described (Giguere

& Prescott, 1998). Furthermore, primers for equine cytokine mRNA have been constructed for

use in "Realtime" PCR assays (Giguere & Prescott, 1999). In the same study, Giguere reports on

the increased sensitivity of Realtime methods over that of conventional PCR.

In the current study, HI and SRH assays are used as reference assays during the

standardization of ELISA techniques. To evaluate the diagnostic performance of the ELISA to

detect serum and nasal antibodies, methods were standardized to grow and purify EIV, determine

the sensitivity and specificity, and develop sampling techniques. Due to various protocols

reported elsewhere, methods to store, culture, and stimulate equine peripheral blood mononuclear

cells (PBMC) were optimized. In addition, Realtime PCR methods were developed to

investigate the expression of equine cytokine mRNA resulting from in-vitro stimulation of PBMC

with EIV antigen.

Materials and Methods

HI Assay

Hemagglutination inhibition (HI) assays were completed as previously described

(Hsiung, 1994). The virus stock used for hemadsorption of test sera was Kentucky-95. Positive

and negative control reference sera were obtained from the National Veterinary Services

Laboratory (NVSL) (Ft. Collins, CO). Allantioc fluid containing live virus was treated with

diethyl-ether at a 1:1 ratio for 15 min. The aqueous portion of the extract was removed and used

in test-wells of a v-bottom, 96-well plate. At the end of the last incubation period, plates were

tilted at a 700 angle and positive inhibition was determined by the absence of "streaming" of

chick RBC in the test wells. Hemagglutination inhibition titers were reported as the number of

hemagglutinating units (HAU) per 25 Al of serum. A 1:128 HI titer indicates a dilution factor of

128 to obtain 1 HAU.

SRH Assay

The measurement of serum antibody levels to EIV was conducted using single radial

hemolysis techniques as adapted from the European Pharmacopoeia Commission. Sheep red

blood cells (SRBC) were sensitized with EIV Kentucky-95 at an HA titer of not less than 1:64 per

25 tl. Guinea pig complement was added to sensitized SRBC which were then added to 1%

agarose at 420 C. The mixture was poured onto plastic plates (ICN Biomedical Inc., Costa Mesa,

CA) marked with mm increments for measuring the zone of hemolysis. Once plates had cooled

to room temperature, 3-mm-diameter holes were punched into the gels and the plates were stored

at 40 C until used. Plates were used not more than three dy after they were made. A negative

control plate was made as described without the addition of virus.

Sera were heat-inactivated at 560 C for 30 min. Ten pl of heat-inactivated test serum was

pipetted into the 3-mm wells and incubated for 20 hr at 340 C. After the incubation period, the

diameter of the clear circular area indicating hemolysis was measured. The zone of hemolysis

was calculated and expressed as mm2 using the following formula:

Zone of hemolysis = ir2 7mm2

where r = radius of the hemolyzed circle.

ELISA Standardization Techniques

Virus purification for use in ELISA

Virus was purified by a procedure adapted from a previously published technique (Laver,

1969). Kentucky-95 EIV was grown in ten-dy-old embryonated chick eggs for 72 hr at 340 C.

Allantoic fluid containing EIV at an HA titer of 1:256 was harvested from eggs and clarified at

4,500 rpm for 30 min at 40 C. EIV was loaded into Beckman ultra-clear centrifuge tubes and

pelleted by centrifugation in a fixed angle (70.1 Ti, Beckman) rotor at 50,000 x g (27,000 rpm)

for 90 min at 4 C. The "slurry" pellet at the bottom of each tube was re-suspended in 50 p~ of

cold TSE buffer and held on ice until pooled. Virus was pooled in a total of 3 mL of cold TSE

buffer. 3 mL of the concentrated EIV (1:1024 HA) was layered onto a 30/%-60% discontinuous

sucrose gradient in Beckman ultra-clear centrifuge tubes, placed in a pre-chilled swing-bucket

(41-Ti, Beckman) rotor, and centrifuged at 95,000 x g (30,000 rpm) for 120 min at 40 C. After

centrifugation, a opaque cloudy band of virus was visualized by holding the tube against a black

background. A 22-ga needle on a 3-mL sterile syringe was used to puncture the side of the tube

and aspirate the cloudy material. To inactivate the virus prior to use on ELISA plates, virus was

adding to Sarcosyl at a final dilution of 0.5% in TSE buffer. The virus and sarcosyl was

incubated for 15 min on ice. The virus was then washed by further dilution in 9 mL of TSE and

centrifuge as previously described at 35,000 rpm for 90 min at 40 C. The purified, inactivated

virus was re-suspended in cold coating buffer containing 0.1% final concentration of sodium

azide (NaN3).

ELISA development

Immunlon-2 HB, 96-well microtiter plates (Dynex Technologies, Chantilly, VA) were

coated with gradient-purified EIV Kentucky-95 diluted in a coating buffer (see appendix B) to 10

HAU per well and incubated at 40 C overnight. The plates were washed 3 times with a wash-

buffer (see Appendix B). The plates were blocked with a 1% fish-gelatin blocking buffer (see

Appendix B) for 2 hr at 370 C, washed 3 times in wash buffer, and stored at 40 C until used. All

test sera, nasal secretions, and reagents (anti-immunoglobulins) were diluted in a 1% fish-gelatin

buffer. Test and control sera were diluted at 1:1000 in fish-buffer. Prior to testing, nasal

secretions were treated with a 1:1 mixture of 10 mM 1-4 dithiotreatole (DTT) (Sigma Chemical

Co., St. Louis, MO) for one hr at 370 C to eliminate non-specific binding of mucous proteins to

the test wells. The nasal secretions were then diluted to a final concentration of 1:300 in 1% fish

buffer. Positive and negative controls were diluted at the same ratio for each test sample.

Positive-reference-serum was obtained from NVSL in Ames, IA. Pathogen-free herd serum was

used as negative-control sera (Sigma Chemical Co., St. Louis, MO). A positive control for IgA

was obtained by collecting nasal secretions from a horse that had been experimentally infected

with live EIV Kentucky-95. This horse was determined to be serum-negative to EIV Kentucky-

95 and 94, Newmarket-96, and Prague-56 prior to infection and serum-positive to Kentucky-95

afterwards as determined by HI assay. Nasal secretions from horses that were serum-negative by

HI assay for EIV Kentucky-95 and 94, Newmarket-96, and Prague-56 were used as a negative

control sample source.

Fifty pl of test samples and control were added to triplicate wells of a 96-well plate.

Plates were sealed with an adhesive plate cover and incubated overnight at 40 C. For serum

samples, plates were washed three times and 50 il of a mouse, anti-horse IgGa (CVM-45)

monoclonal antibody diluted to 1:40 in fish buffer was added to each well. Plates were incubated

for two hr at 370 C. Plates were washed 3 times and 50 p1 of secondary goat anti-mouse IgG +

IgM H+L chains (minimal cross-reactive to human, bovine and horse serum proteins) (Jackson

ImmunoResearch Laboratories, Inc., West Grove, PA) diluted in fish buffer to 1:5000 was added

to each well. After a two-hr incubation at 370 C, the plates were washed prior to the addition of

100 pl O-phenylenediamine dihydrochloride (OPD) (Sigma Chemical Co., St. Louis, MO) to

each well followed by 50 pl of a 3M HCI stop solution. Color development took approximately

ten min. The plates were read immediately at 490 nm using a Revolution, ELISA plate reader

(Dynex Technologies Inc., Chantilly VA). Nasal secretion samples were handled in the same

manner as the serum samples. However, the primary antibody anti-horse IgA (BVS-1) was

diluted at 1:40 and the secondary antibody was diluted at 1:2000 in fish buffer. Results were

reported as an ELISA index.

IgGa (CVM-45) and IgA (BVS-1) were extensively characterized during the Second

International Equine Leukocyte Antigen Workshop, Squaw Valley, California, July 1995. The

specific workshop antibody number is indicated in parentheses (Lunn, Holmes et al., 1998).

To standardize the test results between plates, the ELISA data were reported as an

"ELISA Index" calculated by using the following formula:

ELISA Index = OD reading of test serum/OD reading of positive control serum

Sensitivity and Specificity of the ELISA

The "critical value" or cut off point, where an animal is considered sero-negative to EIV,

is based on criteria that are related to the sensitivity and specificity of the ELISA under

standardized conditions established in our laboratory.

The cutoff value was determined for the ELISA using methods previously described.

Briefly, Immulon II HB plates (Dynex Technologies Inc.) were coated with 10 HAU per well of

gradient purified Eq/2/Ky/95 EIV. Standardization and optimization of reagents were completed

by adjusting dilutions of control and test sera (1:1000), primary antibody against equine IgGa

(1:40), and peroxidase-conjugated secondary antibody (1:5000). Serum samples from 30 horses

that were either seronegative or those that had been experimentally infected and had antibodies to

EIV were used. Serological evidence of exposure to EIV, as determined by HI and SRH assays

and a known history of experimental infection, were used as inclusion criteria for samples. Sera

from horses that had been infected by aerosolized chick allantoic fluid containing approximately

106-" EIDso and had HI titers higher than a 1:64 dilution and SRH titers higher than 75 mm2 were

used as positive controls. Sera with no known previous exposure to EIV or vaccines to EIV and

had HI titers of less than 1:8 and no detectable SRH titers were used as negative controls.

Samples were randomly ordered and labeled using coded numbers generated from an

Excel software program. Triplicate wells containing positive and negative samples were run on

two, 96-well micro-titer plates on the same day. Plates were treated as described above and OD

readings were collected on test samples using an ELISA reader at 490 nm.

Data was used in 2 x 2 tables to determine the sensitivity and specificity of the ELISA at

various "critical or cut-off values". The critical value represents the OD reading generated by the

ELISA that correlates with seroconversion. Sensitivity and specificity were confirmed and

represented constructing a "Critical-Value" plot similar to that used for Receiver-Operating

Characteristic (ROC) plot analysis (Zweig & Campbell, 1993). In the present study, a decision

threshold represented the point at which the ELISA was able to detect antibody in serum samples.

This is referred to as the "critical-value" and was used to establish a reference value at which

horses were considered to seroconvert after vaccination or infection.

Lymphocyte Proliferation Assay

Proliferation assays were adapted from methods described elsewhere (Ellis, Bogdan, &

Kanara, 1995;Hannant & Mumford, 1989;Hannant, 1994). PBMC were collected and isolated

from whole blood using methods described elsewhere (see sample collection, chapter 4 ). Frozen

PBMC were thawed rapidly in a 370 C water bath and washed twice by centrifugation at 250 x g

for ten min at 40 C in PBS. The cells were re-suspended in warm RPMI-1640 with 10-mM Hepes

buffer, 2-mM L-glutamine, 0.075% w/v sodium bicarbonate, 1 mM sodium pyruvate, 100 U/mL

penicillin-G sodium, 100 gg/mL streptomycin, 10% BFS (Gibco BRL, Grand Island, NY), and

recombinant human Interleukin-2 (rh IL-2) at varying concentrations Units/mL (Sigma Chemical

Co., St. Louis, MO), and then cultured overnight before antigen and mitogen stimulation. Cells

were adjusted to 1.5 x 106 cells/mL and added to a 96-well, U-bottom plate at 100 pl per well.

Proliferation assays were conducted to determine the effects of varying the concentration of rh

IL-2, and the infectious-dose of EIV. Blastogenesis was measured in response to heat-inactivated

(570 C for 30 min), UV-inactivated (Newton, Wood et al., 1999), and live

A/equine/2/Kentucky/95 and 91 EIV. The A/2/Kentucky-91 strain of EIV (contained in the MLV

vaccine) was cultured in ten-dy-old embryonated chick eggs (HA titer of 1:64) and frozen at -800

C, prior to its use in the proliferation assays. To compare the response of PBMC stimulated with

live EIV to those stimulated with inactivated antigen preparations, cells were cultured with live

virus (EIDso 1083) for 45-min, washed two times, and resuspended in CM. The cells were

cultured for four dy at 370 C and 5% CO2 Positive controls included stimulation with

Concanavalin A (Con A) (Sigma Chemical Co., St. Louis, MO) at 5 pg/mL and pokeweed

mitogen (Gibco BRL. Grand Island, NY) at 4 pg/mL. Negative controls included media alone

and allantoic fluid without virus. Cells were incubated at 370 C at 5% CO2 for 96 hr. The antigen

specific and mitogen induced proliferation response was measured by using the Cell Titer 96T

Non-Radioactive Proliferation Kit (Promega, Madison, WI). This method is based on the

addition of a tetrazolium salt dye ([3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-

(4-sulfolphenyl)-2H-tetrazolium (Owen's reagent) that is subsequently converted into a colored

formazan product detectable by an ELISA reader. At the end of a four-hr incubation at 370 C

and 5% CO2. Color development resulting from the formazan product was measured at 570 nm

using a Revolution ELISA plate reader (Dynex Technologies, Chantilly, VA). A stimulation

index was calculated as follows where the un-stimulated cell control represents cultures incubated

in media alone:

Stimulation Index (SI) = Absorbance of stimulated cells Absorbance of medium
Absorbance of un-stimulated cells Absorbance of medium

To confirm the sensitivity of the ProMega Cell Titer 96TM Non-Radioactive Proliferation

Kit, duplicate cultures of PBMC stimulated with Con A and EIV antigen were assayed to

determine the proliferation response using radioactive [3H] thymidine uptake as well as the

tetrazolium dye method. To determine potential changes in the viability of equine PBMC after

freezing, cells were assayed as described above prior to and after cryopreservation in liquid

nitrogen. Prior to viability staining and the addition of mitogen or EIV, frozen cells were thawed

and allowed to incubate overnight in three-mL, Falcon polyethylene tubes (Becton Dickinson,

Franklin Lakes, NJ) at 370 C and 5% CO2. This allowed the leaching of DMSO used during


Cytokine mRNA Determination in Equine Peripheral Blood Mononuclear Cells
RNA isolation and reverse transcription

Prior to viability staining and the addition of EIV antigen, frozen cells were thawed and

allowed to incubate in three-mL, Falcon polyethylene tubes (Becton Dickinson, Franklin Lakes,

NJ) overnight at 370 C and 5% CO2 Lymphocytes were cultured in 24-well plates at 5 x 106 to 1

x 107 cells/well. The cultures were stimulated with 128 HAU of heat-inactivated EIV Kentucky-

95 and incubated for 72 hr at 370 C and 5 % CO2. Afterwards, cells were washed twice in PBS

and total RNA was isolated by the use of a Rneasy Mini Kit (Qiagen, Valencia, CA). All RNA

samples were treated with amplification grade DNAse I (Gibco BRL, Rockville, MD) to remove

any traces of genomic DNA contamination. Briefly, 1 U of DNAse I and 1 il of 10 X DNAse I

reaction buffer were mixed with 1.5 utg of total RNA in a ten pl reaction. The mixture was

incubated for eight min at room temperature and then inactivated by adding one .1 of 25 mM

EDTA and heating at 65 OC for ten min.

cDNA was synthesized with a Clontech 1st strand cDNA synthesis kit (Clontech, Palo

Alto, CA) by the method recommended by the manufacturer. Briefly, one pg of total RNA

was mixed with one pi of oligo (dT)ig primer (20 pM) and heated at 700 C for two min. After the

mixture was cooled to room temperature, the following reagents were added in the order listed:

four mL of 5X reaction buffer (250mM Tris-HCL [pH 8.3], 375 mM KCL, and 15 mM MgCI2),

one pl of deoxynucleoside triphosphates (10 mM each), 0.5 pi of Rnase inhibitor (40 U/jl), and

one pC of Moloney murine leukemia virus reverse (MMLV) (200 U/pl). The mixture was

incubated at 420 C for one hr, heated at 94C for 5 min, diluted to final volume of 100 pi, and

stored at -700 C until used for PCR analysis.

cDNA in reverse transcription products was confirmed as previously described (Giguere

& Prescott, 1998). Briefly, two p1 of cDNA was amplified using primer pairs for equine P actin

(Clontech, Palo Alto, CA) in a 25-pi PCR assay in the presence of 0.4 JpM of each primer, 0.2

piM (each) ofdeoxynucleoside triphosphates, five p1 of 10X reaction buffer (containing 10 mM

TrisHCl [pH 8.3] and 50 mM KCI), 1.5 mM MgCl2, and two U of Taq DNA polymerase

(AmpliTaq; Perkin-Elmer, Branchburg, N. J.). Polymerase chain reactions were performed with

an initial denaturation step at 940 C for 2 min and 35 cycles of amplification followed by a 7-min

extension at 720C. Each cycle included denaturation at 94C for 45 s, annealing at 600C for 45 s,

and extension at 720C for 2 min. Amplified PCR products were visualized by gel electrophoresis

of ten .1 of product on a 1.6% agarose gel followed by ethidium bromide staining for 15 min. The

specificities of the amplified bands were confirmed by their predicted product size based on a

molecular weight standard. Positive (24-hr Con A stimulated equine PBMC) and negative

(reaction master-mix only) controls were added to identify specific and non-specific amplification

during PCR.

Real-time PCR utilizes the 5' to 3' endonuclease activity of AmpliTaq GoldTM DNA

Polymerase to cleave a TaqMan probe. During PCR, if the target molecules are present, the

probe specifically anneals between the forward and reverse primer sites. The TaqMan probe

contains a reporter dye (6-carboxyfluoresceine) at the 5' end and a quencher dye (6-

carboxytetrmethylrhodamine) at the 3' end. The un-reacted 3' nucleotide is blocked by

phosphorylation, preventing elongation of the Taq DNA polymerase and dramatically reducing

the flourescence of the reporter dye. The endonuclease activity at the 5' end of AmpliTaq Gold

causes the probe to be cleaved allowing the separation of the reporter and quencher dyes resulting

in an increased flourescence of the reporter dye. The accumulation of PCR products was detected

by an ABI Prism 7700 Sequence Detector (Applied Biosystems, Foster City, CA). The amount

of cytokine mRNA in test samples was expressed as a fold difference compared to that seen in

equine PBMC after 24-hr Con A stimulation (positive control). The control sample provided a

source of cDNA with an adequate expression of cytokine mRNA that could be used as a standard

for each PCR reaction. Cytokine mRNA expression in the cDNA standard was used to compare

to that seen in test samples and subsequently reported in fold-differences.

Real-time PCR analysis

All real-time PCR reactions were performed in special optical tubes that focus the

flourescence signal in a 96-well format. PCR reactions were completed in an ABI PRISM 7700

Sequence Detector System (Perkin Elmer, Foster City, CA). Forward and reverse primers (900

nM/pl) for target cytokine (IL-2, INF-y, IL-4, and IL-6) sequences and TaqMan probes (250

nM/pl) (Applied Biosystems, Foster City, CA) were reacted with 12 pJ of TaqMan Universal

PCR Master Mix, 4.5 pl of DEPC treated water, and 2 p.1 of sample cDNA per tube. The internal

probes were labeled at the 5' prime end with the reporter dye 6-carboxyfluoresceine, and at the 3'

prime end with the quencher dye 6-carboxytetrmethylrhodamine. Fluorescence signals were

generated during each PCR cycle by the 5'=> 3' endonuclease activity of AmpliTaq GoldTM.

Amplification was performed with initial incubation steps at 50C for two min and 95C for ten

min followed by 40 cycles of 95C for 15 min and 60C for one min. All samples were assayed

in triplicates and the mean values used for comparison. To account for variation in the amount

and quality of starting material, all the results were normalized to G3PDH expression. Relative

quantitation between samples was achieved by comparing their normalized threshold cycles (Ct).

The Ct represents the PCR cycle at which an increase in reported flourescence above the

threshold is detected. Samples without cDNA were included in the amplification reactions to

determine background flourescence and check for contamination. cDNA from 24-hr Con A-

stimulated equine PBMC was used as a positive control. The level of mRNA expression was

reported as an "X"-fold difference to the calibrator. The samples used in the Realtime PCR assays

were those collected during animal studies to investigate the immune response of horses to

vaccination and infection with EIV.



The HI and SRH tests were used to screen horse sera for the presence of antibodies to

various strains of EIV prior to their enrollment. The HI and SRH were also effective at detecting

rises in serum antibody levels after vaccination and infection. Sera from day zero through 70 of

the study were assayed by the HI and SRH method. While the HI assay was effective in

establishing an antigen-specific antibody response in horse sera, the SRH test was found to be

more sensitive than HI in detecting the appearance of circulating antibodies to EIV at an earlier

time point after vaccination or infection. The sera from horses infected with the Kentucky-95

strain of EIV reacted with both Kentucky-95 and 94 antigen used the HI assay.

The ELISA has been used previously to detect isotype-specific immunoglobulin

in equine sera and nasal secretions after vaccination and infection. In the present study, the

ELISA provided a reliable method to measure serum IgGa and local IgA antibodies following

vaccination and infection. The ELISA was used to screen horse sera for the presence of antibody

prior to their enrollment. Hemagglutination and SRH tests were completed on the same sera to

confirm the specificity of the ELISA. The sensitivity and specificity of the ELISA were further

investigated and tests were conducted that provided a method to determine a "critical value" or

cut-off value representing seroconversion. The OD at 490 nm was determined for each sample by

an ELISA reader (Figures 2-1 and 2-2).

I I I I I l l i t t i t t i l i t a t l I l t i l l I l l

2 3 9 10 11 121316 171819 20 21

23 24 28 34 35 37 38 40 41 42 44 45 55 56 58 59 60

Animal ID

Figure 2-1 ELISA optical density readings from 30 horses that were confirmed as serum positive
by hemagglutination inhibition and single radial hemolysis tests.




E 0.1

ii 0.06
0 0.04



1 5 6 7 8 1415 2225 262729 30313233 36 3943 464748 49 505152 535457

Animal ID

Figure 2-2. ELISA optical density readings from 30 horses that were confirmed as serum negative
by hemagglutination inhibition and single radial hemolysis tests.
Note: Scale is 1/10 of previous graph.

Determination of the Critical or Cut off Value for the ELISA

Disease Status

ELISA Results

Sensitivity = a/a+b
Specificity = d/c+d

Figure 2-3.2 x 2 tables used to calculate the sensitivity and specificity of The ELISA.

2 x 2 tables using data from Figures 2-1 and 2-2 were constructed to calculate the percent

sensitivity and specificity at various OD readings (Figure 2-3). The ELISA was sensitive enough

to detect antibodies in the sera of horses that were truly positive and distinguish from those that

were truly negative. Zero of 30 sera in the seropositive group had an OD less than 0.1 and only

two of 30 were less than 0.2. Further, only one serum sample from the seronegative group had an

OD greater than 0.1. Table 2-1 summarizes the relative sensitivity and specificity at various

critical values ranging from 0.05 to 1.0. Based on the data summary in Table 2-1, an OD reading

of 0.1 correlated with the highest sensitivity and specificity.

a b

c d

Table 2-1 Critical (Cut-off) Values and the Associated Sensitivity and Specificity for the ELISA.
Sensitivity and specificity at different cut-offvalues was determined using 2 x 2 tables described
in figure 2-3.

Cut-off Values















% Sensitivity















% Specificity















Generating a "Critical-Value" Plot

A critical-value plot was constructed to show changes in sensitivity and specificity at

various critical-values (Figure 2-4). As the critical value decreases, inclusion of a true-positive

test-result produces an increase in the vertical trend reflecting higher sensitivity. Inclusion of a

false-positive test-result produces a horizontal line. The point on the curve representing the

critical value (cut-off value) which lies closest to 100% for both sensitivity and specificity is

indicated by the plot and was used to assess acute and post-inoculation sera. Based on these

data, OD readings greater than 0.2 were considered to be representative of horses that were

positive for EIV-specific antibody. 'The ELISA provided a reliable quantitative method to detect

serum IgGa antibody with minimal variability (standard deviation typically less than 0.1) between

triplicate wells.


(True Positives)




Figure 2-4. A Critical-Value plot of serum IgG antibody detected in seropositive and seronegative
horses. Sensitivity and specificity for the ELISA are graphically represented by plotting the
number of false positive (y axis) against the true positives (x axis).

Lymphocyte Proliferation Assay

Peripheral blood lymphocytes from horses in each group were used to determine the

antigen-specific response to in-vitro stimulation with heat-inactivated A/equine/2/Kentucky/95


* *(0.o5)

EIV in a 96-hr proliferation assay. Figure 2-5 indicates that the viability of cryopreserved equine

PBMC averaged 89% upon thawing. Incubation of cultures for 18 hr to leach out

cryopreservation reagents reduced the viability to 68%. Further, the process of freezing, thawing,

the associated washes and centrifugation resulted in a 44% loss of cells. Con A was used to

confirm the functional capacity of lymphocytes. Con A had a dose-dependent response in fresh

and frozen equine PBMC. The optimum concentration of Con A was determined to be 5 pg/mL

(Figure 2-6) and was used during the remainder of the study. A comparison of two assay

methods to detect the blastogenic response of Con A-stimulated cells was performed. The

blastogenic response, measured by a [3H] Thymidine uptake assay, was comparable to that of the

tetrazolium dye method (Figure 2-7). The stimulation index did not differ appreciably in either

assay method. A SI greater than or equal to two was considered a positive antigen-specific

proliferation response. All PBMC cultured during the blastogenesis assays were found to be

immunologically functional by stimulation with either Con A or Pokeweed mitogen (data not


Various preparations (heat-inactivated, UV-inactivated, and live) of EIV were used to

determine if differences existed in the in-vitro response of equine PBMC. Based on the results

represented in figure 2-8 and other experiments (data not shown), heat inactivated EIV was

determined to induce a measurable blastogenic response in equine PBMC collected from horses

following infection. Other assays using UV-inactivated virus demonstrated greater variability in

the response (data not shown). No appreciable increase in the proliferation response was noted

by using live virus compared to heat or UV-inactivated preparations as well.

Figure 2-9 shows a subtle change in the proliferation response from a range of infectious

doses of EIV. Based on these results and other suggested protocols, an EIDo 1083 was used in

subsequent assays. In the next experiment, the dose-dependent response of PBMC to hrIL-2 was

assayed using EIV antigen and Con A mitogen. Figure 2-9 indicates a subtle difference in the

Con A-stimulated response resulting from increasing concentrations of hrIL-2 in the CM.

However, a dose response was seen in PBMC stimulated with EIV when varying the

concentration of hrIL-2. The stimulation index decreased when the hrIL-2 concentration was

increased. In contrast, un-stimulated PBMC appeared to benefit from increased concentrations of

hrIL-2. Based on these data, culture media was supplemented with 20 U/mL of hrIL-2 and used

in subsequent proliferation assays.


r -
* 60
U 50
. 40-

Viable PBMC Thawed

Viable PBMC After Overnight

Figure 2-5. Changes in the total cell number recovered after thawing and viability after overnight
incubation of cryopreserved equine peripheral blood mononuclear cells. The left hand-side bar
reflects the total number of cells remaining after freezing, thawing, and washing. The center bar
represents viability of cryopreserved equine peripheral blood mononuclear cells that were rapidly
thawed in a 37 C water bath. Viability was determined by a trypan blue exclusion-dye method.
The right hand-side bar shows a reduction in viability of frozen/thawed equine peripheral blood
mononuclear cells after an 18-hr incubation at 370 C and 5% CO2. Error bars represent SE of
the mean.

PBMC Recovered After






10 5 2.5 1.25 0.625 0.312 0.158 0.078 0.039 0.019 Cells

micrograms/ml of Con A

Figure 2-6. Dose-dependent response of equine peripheral blood mononuclear cells (1 x 105
cells/ml) to Concanavalin A during a 72-hr proliferation assay. Freshly isolated equine peripheral
blood mononuclear cells were stimulated in vitro under culture conditions described in materials
and methods. The mitogen-induced proliferative response was measured using a colormetric,
tetrazolium dye assay method, also described in materials and methods. Data is representative of
three separate experiments.

10 --

Con A

Live Virus

* PHI ThymMine
DTetrazollum Dye

HI Virus

Antigen/Mtogen Stimulation

Figure 2-7. Comparison of [3H] Thymidine and a tetrazolium dye colormetric assay methods to
detect equine lymphocyte proliferation. Equine peripheral blood mononuclear cells were
stimulated in vitro with Concanavalin A at 5 p.g/mL for 72 and the proliferative response was
then measured by both [3H] thymidine incorporation and a tetrazolium dye colormetric assay.
Data is representative of individual assays from three different horses.









6 ----------------------------------


UV HI Live Con A
Antigen Preparation

Figure 2-8. Comparison of various methods of antigen preparation (heat inactivation, UV
inactivation, and live virus) for in-vitro stimulation of peripheral blood mononuclear cells with
equine influenza virus. Equine influenza virus was inactivated by UV light using a technique
described elsewhere {Rott & Cash 1994 732 /id}, or heat inactivated at 570 C for 30 min. Cells
were stimulated with live virus at an infectious dose of 1083 EID50. Concanavalin A was used at
a concentration of 5 pg/mL. Cells were isolated from six horses, eight days after infection and
were cryopreserved prior to the proliferation assay. Error bars represent SE of the mean.


X 6


-- --





Antigen/Mitogen and IL-2

Figure 2-9. The in-vitro dose-dependent response of equine peripheral blood
mononuclear cells to varying amounts of EIV and human recombinant IL-2
supplementation was determined in peripheral blood mononuclear cells isolated from
horses 14 days after infection with EIV. The proliferative response of cells to varying
amounts of equine influenza virus was determined with rhIL-2 supplemented to the
culture medium as indicated. In a separate series of experiments, the dose of EIV was
held constant and the effect of increasing amounts of rhlL-2 was determined. The
response of equine peripheral blood mononuclear cells to increasing concentrations of
rhIL-2 was then determined in both mitogen stimulated and un-stimulated cultures, as

i" Ill.
m m m m m m m
do do o do o o o
a0 0 00

Sr r 5
&&&F&& I




Cytoldne mRNA Determination in Cryopreserved Equine PBMC

Reverse transcription of RNA was followed by conventional PCR methods to confirm P3-

actin mRNA in PBMC stimulated in vitro with EIV for 72 hr. Figure 2-10 is representative of

several experiments showing the relative expression of 13-actin in cDNA samples using rt-PCR

assays. After p1-actin was confirmed in each sample, cDNA was analyzed for G3PDH and

cytokine mRNA expression by Realtime PCR. The relative expression of G3PDH was found to

be comparable to levels expressed in Con A-stimulated cells. Total cDNA was then analyzed for

IL-2, INF-y, IL4, and IL-6 mRNA expression. Whereas levels of IL-2, INF-y, and IL-6 were

readily detected in most of the samples, IL-4 was expressed at very low levels. There were no

clear patterns of cytokine expression in samples from vaccinated and EIV-infected horses.

Furthermore, cytokine mRNA from in-vitro stimulated equine PBMC was highly variable and did

not show a pattern of vaccine-induced expression (Figures 2-11 through 2-14).

It was speculated that a 72-hr stimulation time was too long to detect peak levels of

cytokine mRNA expression in cultured PBMC. Therefore, a second experiment was conducted

to determine the relative expression of IL-2 mRNA in PBMC from two vaccinated horses after an

8-hr stimulation. An increase was noted in both samples after an 8-hr in-vitro stimulation

(Figures 2-15 and 2-16). As indicated in the graphs, the pattern of IL-2 expression was reversed

in the 8-hr culture compared to that seen after 72-hr in-vitro stimulation. These data indicate that

frozen PBMC from horses can be successfully assayed for cytokine expression using Realtime

PCR methods. Based on these results, subsequent cytokine analysis was conducted in 8-hr in-

vitro stimulation cultures.

LD 114 119 120 114 119 120 148 150 ConA Neg

800 bp

628 bp

400 bp

S PIDo } PID 7 }) PID 0)

Figure 2-10. Amplified PCR products representing p actin (628 bp size fragment) mRNA
expression visualized by gel electrophoresis and ethidium bromide staining. Equine peripheral
blood mononuclear cells were stimulated in-vitro, total RNA was isolated, and cDNA was
generated by reverse transcription. cDNA was confirmed by a conventional PCR assay using
primers for p actin. These data are representative of mRNA isolated from in-vivo-primed cells
collected at various time points (post inoculation days [PID] 0 and 7) after vaccination and
infection with equine influenza virus.

200 -







- MLVVacc
Inact Vacc
-- Control
--f- Recomb Vacc

Vacc 7 2nd Vacc 7 Chall 7
Sample Days

Figure 2-11. IL-2 mRNA expression in in-vivo primed peripheral blood mononuclear cells from
vaccinated and infected horses during a 72-hr in-vitro stimulation with heat-inactivated equine
influenza virus antigen. Equine peripheral blood mononuclear cells were stimulated in-vitro, total
RNA was isolated, and cDNA was generated by reverse transcription. A Realtime PCR assay was
conducted to show changes in mRNA cytokine expression at various times (day of vaccination
[Vacc], 7 days after vaccination or challenge [Chall]) as indicated. Cytokine detection assays
were not done on 2nd Vacc or Chall days. Changes are expressed as a x-fold difference above a
Con A control. (Vaccines and EIV-INF groups n = 3; Control n = 1)



7 2nd Vacc 7
Sample Days


Figure 2-12. INF-y mRNA expression in in-vivo primed peripheral blood mononuclear cells from
vaccinated and infected horses during a 72-hr in-vitro stimulation with heat-inactivated equine
influenza virus antigen. Equine peripheral blood mononuclear cells were stimulated in-vitro, total
RNA was isolated, and cDNA was generated by reverse transcription. A Realtime PCR assay was
conducted to show changes in mRNA cytokine expression at various times (day of vaccination
[Vacc], 7 days after vaccination or challenge [Chall]) as indicated. Cytokine detection assays
were not done on 2nd Vacc or Chall days. Changes are expressed as a x-fold difference above a
Con A control. (Vaccines and EIV-INF groups n = 3; Control n = 1)


! 250






-MLV Vac
ca nact Vacc
-- Control
-f-Recomb Vacc




i -llMLV Vacc
SC- Inact Vacc
5 -*-Control
S--a-Recomb Vacc

so L---- -

Vacc 7 2nd Vacc 7 Chall 7

Sample Days

Figure 2-13. IL-4 mRNA expression in in-vivo primed peripheral blood mononuclear cells from
vaccinated and infected horses during a 72-hr in-vitro stimulation with heat-inactivated equine
influenza virus antigen. Equine peripheral blood mononuclear cells were stimulated in-vitro, total
RNA was isolated, and cDNA was generated by reverse transcription. A Realtime PCR assay was
conducted to show changes in mRNA cytokine expression at various times (day of vaccination
[Vacc], 7 days after vaccination or challenge [Chall]) as indicated. Cytokine detection assays
were not done on 2nd Vacc or Chall days. Changes are expressed as a x-fold difference above a
Con A control. (Vaccines and EIV-INF groups n = 3; Control n = 1)



so -
I MLV Vacc
Se Inact vacc
60 o- BEIV-4NF
0I -&-Control
0 -a-Recomb Vacc


0- L
Vacc 7 2nd Vacc 7 Chall 7
Sample Days

Figure 2-14. IL-6 mRNA expression in in-vivo primed peripheral blood mononuclear cells from
vaccinated and infected horses during a 72-hr in-vitro stimulation with heat-inactivated equine
influenza virus antigen. Equine peripheral blood mononuclear cells were stimulated in-vitro, total
RNA was isolated, and cDNA was generated by reverse transcription. A Realtime PCR assay was
conducted to show changes in mRNA cytokine expression at various times (day of vaccination
[Vacc], 7 days after vaccination or challenge [Chall]) as indicated. Cytokine detection assays
were not done on 2nd Vacc or Chall days. Changes are expressed as a x-fold difference above a
Con A control. (Vaccines and EIV-INF groups n = 3; Control n = 1)





Vacc 7


Vscc 7
Sample Days

Figure 2-15. IL-2 mRNA expression in peripheral blood mononuclear cells from horse # 146 in
the modified-live vaccine group. Equine peripheral blood mononuclear cells underwent in vitro
stimulation for 72 or 8 hr; total RNA was isolated, and cDNA was generated by reverse
transcription. A Realtime PCR assay was conducted to detect changes in mRNA cytokine
expression on the day of and 7 days following vaccination (Vacc) as indicated. Changes are
expressed as a x-fold difference above a Con A control.

m l--------------------



100 -

0 --

*M72 hr

Sample Days

Figure 2-16. IL-2 mRNA expression in peripheral blood mononuclear cells from horse # 149 in
the inactivated virus vaccine group. Equine peripheral blood mononuclear cells underwent in
vitro stimulation for 72 or 8 hr; total RNA was isolated, and cDNA was generated by reverse
transcription. A Realtime PCR assay was conducted to detect changes in mRNA cytokine
expression on the day of and 7 days following vaccination (Vacc) as indicated. Changes are
expressed as a x-fold difference above a Con A control.


Historically, HI, SRH, and virus neutralization assays have been the gold-standard for

measuring sero-conversion to vaccination or infection with EIV(Wilson, 1993) (Mumford, 1991).

These assay methods have been shown to correlate closely with virus neutralizing antibodies

(Morley, Hanson, Bogdan, Townsend, Appleton, & Haines, 1995). The HI assay is the most

widely used serological indicator of immunization against or infection with influenza in humans,

mice, and horses. It is relatively inexpensive, rapid, and can differentiate between types 1 and 2

of equine influenza (Wilson, 1993). The specificity of the HI test is based on the principle that it

detects antibodies to the HA surface protein which is subject to antigenic drift. However, in our

hands this assay was not as sensitive as either SRH or ELISA techniques to detect seroconversion

at an early time point after exposure to EIV.

Because a four-fold increase in the HI antibody titer is considered diagnostic of exposure

to influenza antigen, detection of a sub-clinical infection is possible. However, variability

between laboratories, assay technicians, and individual assays has been reported (Mumford,

1991). Some of the variability can be explained by the different methods used to disrupt the virus

prior to its use as antigen. Furthermore, one study showed that antigen from influenza grown in

cell culture (Madin-Darby canine kidney cells) is more sensitive to antibody than that derived

from virus grown in chicken eggs (Cook, Mumford et al., 1988).

The SRH assay is widely accepted as a sensitive and reproducible indicator of sero-

conversion and studies have shown that SRH titers ranging from 75 to 150 mm2 correlate with

protection against infection with EIV (Hannant, Mumford, & Jessett, 1988;Mumford, Wood,

Scott, Folkers, & Schild, 1983;Mumford, Wood, Folkers, & Schild, 1988;Mumford, Jessett,

Dunleavy, Wood, Hannant, Sundquist, & Cook, 1994a;Mumford, Wilson, Hannant, & Jessett,

1994;Townsend, Morley et al., 1999;Wood, Mumford, Folkers, Scott, & Schild, 1983). The

increased sensitivity to detect circulating antibodies to influenza by SRH assay over that of HI

has been reported elsewhere (Ennis, F.A. et al., 1977).

In previous studies, detection of serum antibody to influenza vaccination and infection

was restricted by assay methods that had limited sensitivity, specificity or provided results that

were variable. In this study, ELISA techniques provided both a sensitive and specific assay to

detect serum and nasal influenza-specific antibodies. Sensitivity and specificity are two indicators

of the amount of validity of a diagnostic test. Sensitivity can be calculated from the relation

(a/a+b) in a two by two table (Figure 2-4) and is described as the probability of having a positive

test result from an animal that is truly diseased where "disease" can mean sub-clinical disease and

or infection. Likewise, specificity is a measure of the probability of a negative test result for a

sample from a truly negative animal (d/c+d). By increasing the sensitivity of the test, the number

of cases that have had exposure and are misdiagnosed as negative by the test (false negatives) will

decrease. Appropriately so, a highly specific test will report a small number of positives in the

absence of exposure. These are referred to as false positives. Because there will be a range of

values for both diseased and non-diseased animals (low and high values in each category), there

will be some overlap which results in an inverse relationship between the sensitivity and

specificity of the ELISA measuring antibody titers. This relationship can be adjusted based on a

selected critical value (cut off point). Therefore, the critical value will delineate those animals

considered positive or negative for sero-conversion. Horses exposed to EIV with OD above the

critical value are considered positive and are represented by a in the two by two table. Those

exposed and having values below the critical value are represented by c and are false negatives.

Horses that have not been vaccinated or experimentally infected and have values above or below

the critical value are represented by b and d respectively.

The ELISA method was able to correlate an OD reading or "critical-value" to seropositive

or seronegative status in horse sera. True positive or negative horses (based on known history

and the determination of influenza-specific antibody by HI and SRH) were detected by the

ELISA with 100% and 97% sensitivity and specificity respectively. This type of analysis

provided the basis to establish a cut-off value (OD value) of 0.1, reflecting the point where horses

were considered to have seroconverted. Based on these parameters, this assay was further used in

the current study to determine the serum and local antibody response post-vaccination or


Due to varying proliferation assay methodologies, preliminary assays were conducted to

optimize culture conditions for this study. Antigen preparation and the duration of incubation

were optimized to obtain a measurable response in PBMC from horses prior to and after

challenge infection. In some cases, samples were collected from horses used from ongoing

investigations within our laboratory not directly related to the present study.

To confirm the ability of lymphocytes to respond to mitogenic stimulation, cultures

containing Con A were included in the proliferation assays. Depending on the stimulus, the

proliferation response can involve either T or B cells or both. For example, during mitogen-

induced proliferation, phytohemagglutinin (PHA), and Con A preferentially induce T cells while

PWM will result in both a T and B-cell response (Fletcher, Klimas, Morgan, & Gjerset, 1992).

For antigenic peptides derived from endogenous or viral protein, class I MHC molecules are

involved (Neefjes & Momburg, 1993;Rammensee, Falk et al., 1993). Extracellular antigens are

processed via endosomal compartments, presented by MHC-II molecules, and activate CD4' T

cells (Roitt, 1997).

One published study reports the dose-dependent response of equine PBMC to Con A at an

optimum concentration of 20 pg/mL during a 4-dy incubation (Truax, Powell et al., 1990). In

contrast, the present study demonstrated unfractionated equine PBMC had an optimum four-dy

proliferation response to Con A at 5 pg/mL. Further, the addition of hrIL-2 at 200 U/mL did not

show a significant increase over that seen with 20 U/mL during Con A stimulation. Others have

reported lymphocyte blastogenesis in response to increased concentrations of hrIL-2 during

mitogen stimulation (Hammond, Issel et al., 1998) and the maintenance of long-term cultures of

hrIL-2-dependent PBMC (Stott & Osburn, 1988). However, as might be expected with all non-

human T lymphocytes, the response of equine PMBC to human IL-2 is less than would be seen

by the addition of a species-specific IL-2 (Fenwick, Schore et al., 1988).

The number of animal studies that have included the use of cryopreserved PBMC is

limited. Cryopreserved PBMC have been used in histocompatibility and mitogen-response

experiments in sheep (Stear, Allen et al., 1982), cattle (Kleinschuster, VanKampen et al., 1979)

and in horses (Truax, Powell, Montelaro, Issel, & Newman, 1990). In contrast, the mitogen-

induced response of human lymphocytes after cryopreservation has been well characterized (see

review) (Fletcher, Klimas, Morgan, & Gjerset, 1992). Furthermore, the successful use of human

mononuclear cells after cryopreservation has been demonstrated (Allsopp, Nicholls et al.,

1998;Lamb, Jr., Willoughby et al., 1995). One study reported that frozen/thawed human

lymphocytes not only recover and respond to plant mitogens but showed a reduction in the

variability associated with proliferation assay techniques (Sears & Rosenberg, 1977).

Cryopreserved equine PBMCs have been used in separate studies to determine the CTL activity

against equine herpes-virus (EHV) (Allen, Yeargan et al., 1995;O'Neill, Kydd et al., 1999).

Consistent with other reports, where cryopreserved cells were 95% viable (Truax, Powell,

Montelaro, Issel, & Newman, 1990), equine PBMC used in the present study were 89% viable

immediately following a rapid thaw. The percentage of viable cells fell to 68% after 18-hr

incubation in culture media however. The cause for the loss of cell viability may have been due

to the lack of cytokines. However, the exact cause of cell death was not investigated. Cells that

were viable after over-night incubation, were consistently capable of responding to Con A ,

Pokeweed mitogen and minimally to EIV antigen.

The proliferative response was measured by a colormetric dye method in place of a [3H ]

thymidine uptake assay. Other studies have reported on the ability of a tetrazolium-salt-based

dye to detect the proliferation response in cultured cells (Behl, Davis, Lesley, & Schubert,

1994;Lappalainen, Jaaskelainen, Syrjanen, Urtti, & Syrjanen, 1994;Wong & Goeddel, 1994). A

comparison of the colormetric and the [3H ] thymidine uptake assays has been made elsewhere

(Kitamura, Tange et al., 1989) and in our laboratory for their ability to detect a proliferative

response. Assays conducted in the present study did not indicate that the [H3] thymidine uptake

method was appreciably more sensitive than the colormetric dye method. Furthermore, the

colormetric method eliminated the need for specialized training and equipment associated with

radioisotope use.

In-vitro exposure to influenza antigen will induce T lymphocyte proliferation (Chow,

Beutner et al., 1979). However, these observations are based on the response of PBMC to

inactivated preparations of virus (Chow, Beutner, & Ogra, 1979;Lazar & Wright, 1980). While

one report demonstrates a depressed mitogen response by influenza-infected human lymphocytes,

the proliferation response to the viral infection remained intact (Roberts, Jr. & Nichols, 1989).

Furthermore, human lymphocytes, infected with influenza, can stimulate autologous cells in

short-term cultures (Thompson, Lewis et al., 1973). Studies describing the use of inactivated

stimulator-cells have been conducted as well (Moreno & Lipsky, 1986). Therefore, in the

present study, a comparison between the use of paraformaldehyde-fixed, UV inactivated

(Roberts, Jr. & Nichols, 1989) autologous antigen-presenting cells, or the addition of heat

inactivated virus to PBMC cultures was investigated. The use of fixed EIV-infected cells proved

to be complicated with no demonstrable increase in the proliferation response between cultures

when compared to other methods, including heat or UV-inactivated antigen preparations or live

virus. In some experiments, a higher percentage of cell death occurred in cultures containing UV-

inactivated or live virus (data not shown). In addition, a comparison was made to determine the

optimal incubation time. Others have described three (Wiley & Skehel, 1977) and four-dy (Ellis,

Bogdan, & Kanara, 1995;Hammond, Cook, Lichtenstein, Issel, & Montelaro, 1997) incubation

periods for proliferation assays. Based on observations made in the present study, heat-

inactivated virus was found to have a measurable blastogenic effect on PBMC and was used to

determine the antigen-specific response in 96-hr proliferation assays.

We used real-time quantitative rt-PCR to demonstrate that in-vivo-primed PBMC isolated

from whole blood, could be frozen, thawed, stimulated in vitro, and assayed for cytokine mRNA

expression. Beta actin expression was confirmed from cDNA after reverse transcription of

mRNA by conventional PCR and gel electrophoresis. Once cDNA was determined for each

sample, real-time PCR assays were conducted to detect glyceraldehyde 3-phosphate

dehydrogenase (G3PDH) and target cytokines.

In the present study, mRNA cytokine production after in-vitro stimulation with inactivated

EIV was investigated. PBMC from horses that had either been infected or immunized with one

of three different vaccine preparations were stimulated in vitro for three days and assayed for

cytokine mRNA expression. While one study shows expression of INF-y up to 48 hrs after in

culture, it also demonstrated an early peak-expression of IL-2 and IL-4 with a subsequent decline

after four hr post-stimulation (Kruse, Moriabadi, Toyka, & Rieckmann, 2001). The importance

of the stimulation-time was demonstrated in the present study. Interferon-y levels increased in

PBMC from horses on seven day after the first and second vaccination and infection. In contrast,

there was a dramatic drop in IL-2 levels on day seven after the initial vaccination or infection. A

similar trend was noted in IL-4 and IL-6. While it is conceivable that IL-4 levels may be lower

than IL-2 and INF-y in response to infection, a similar pattern of IL-4 expression was seen in the

inactivated-virus vaccine group and negative control. Based on studies in mice, IL-6 may have

been expected to rise after either vaccination or infection (Matsuo, Iwasaki et al., 2000). A

possible explanation for the highly variable cytokine mRNA expression in the present study is

that the stimulation time of PBMC was too long (72 hrs), and peak expression of INF-y, IL-2, IL-

4, and IL-6 was within the first four to 48 hours after antigen stimulation. These data have not

been published for horses. Therefore, a pilot study to investigate the kinetics of cytokine

expression in PBMC at various times during a 48-hr in vitro stimulation with Con A would have

been appropriate. To make this comparison, however, would have required additional lengthy

animal studies. Speculation as to why mRNA levels at day zero were many-fold higher than at

dy-7 should include consideration of the molecular events regulating the memory cell phenotype.

Based on previous studies already mentioned, an increase in IL-2 mRNA cytokine expression was

expected day seven after inoculation with either live virus or an MLV vaccine. In mice,

differential cytokine expression can be detected within hours of stimulation. However, the data

show a dramatic decline in all groups on day seven compared to values on day zero. The

dramatic decrease in cytokine levels on day seven may be associated with the in-vitro culture

time. By the third day of incubation, the cellular processes responsible for gene expression were

depleted in PBMC responding in an antigen-specific (primed) manner but not those that had no

previous exposure to EIV antigen. This idea is supported by the fact that PBMC on day zero

may have not undergone an increased upregulation of cellular mechanisms (gene expression) in

response to antigen stimulation. Interferon-gamma and IL-2 mRNA from un-stimulated mouse

lymphocytes were shown to reach maximum expression on day seven after influenza infection

(Matsuo, Iwasaki, Asanuma, Yoshikawa, Chen, Tsujimoto, Kurata, & Tamura, 2000) (Mosmann

& Coffman, 1989). Also, murine lymphocytes have been shown to reach maximum expression of

INF-y and IL-2 in day seven post-infection samples after 24-hr in-vitro stimulation (Asanuma,

Aizawa, Kurata, & Tamura, 1998). Furthermore, expression of IL-4 and IL-6 has been shown to

peak in nasal associated lymphocytes of mice seven days after vaccination and infection with

influenza, respectively (Matsuo, Iwasaki, Asanuma, Yoshikawa, Chen, Tsujimoto, Kurata, &

Tamura, 2000). Again, in the Matsuo study, no in-vitro stimulation was used. In other studies,

cytokines have been measured in the supernatants of human macrophage and lymphocyte cultures

after infection with influenza. In humans, INF-y, IL-2, IL-4, and IL-6 activity was quantified in

the supernatants of PBMC post-vaccination by ELISA methods (McElhaney, Upshaw et al.,

1998). In another study, peak protein levels of INF-y have been measured in the supernatants of

mouse macrophage cultures seven days after infection (Monteiro, Harvey et al., 1998).

The hypothesis (relative to the current study) that peak cytokine mRNA expression was

missed in vitro and the related cellular mechanisms were exhausted in cells primed in vivo but not

un-primed cells by the third day in culture is further supported by Figure 2-15. To investigate the

earlier expression ofIL-2 mRNA, PBMC from two horses (# 146 in the MLV vaccine group and

# 149 in the inactivated vaccine group) were stimulated for 8 hr and assayed for cytokine

expression as described above. Day zero IL-2 levels induced by stimulation for eight hours, were

within only a 1.6 fold difference of those resulting from a 72-hr stimulation. However, a rise in

IL-2 mRNA expression in dy-7 samples was noted in the 8-hr stimulation cultures. Furthermore,

overall levels of IL-2 were higher in the 8-hr versus the 72-hr cultures. Again, this is most likely

due to the fact that the PCR assay is able to detect relative mRNA expression at a particular point

in time and is a measure of the in-vivo-primed cell's ability to respond in an antigen-specific

manner. Based on this information, additional tests are ongoing to determine the differential

cytokine mRNA expression in PBMC using an 8-hr in-vitro stimulation assay. It will be helpful

to determine the cytokine response in addition to other measures of immunity in order to fully

characterize the relationships between each.

In summary, it appears that the standardized assay methods described above can be used

to characterize the immune mechanisms in horses. The ELISA could be used to detect small

increases in serum and local antibody with a sensitivity and specificity of 100% and 97%,

respectively. Furthermore, use of rayon-tipped proctoscopic swabs proved to be an easy and

efficient method to collect nasal secretions from horses. Lastly, the above defined culture

conditions are appropriate to assay cryopreserved equine PBMC in 98-hr lymphocyte

proliferation assays and cytokine mRNA expression analysis.



One of the difficulties in the field of equine influenza is the diversity of techniques

associated with challenge-infection models. Specifically, various methods have been used to

challenge/inoculate experimental animals with EIV during vaccine efficacy testing. Methods of

inoculation have included intranasal instillation (Mumford, Wood, Scott, Folkers, & Schild,

1983;Mumford, Wood, Folkers, & Schild, 1988) and aerosolization of live virus into an enclosed

stall by use of a nebulizer (Mumford, Wilson, Hannant, & Jessett, 1994). Nebulizers have been

the more recent choice because they aerosolize virus in droplets with a diameter of less than 5p.

that reaches the upper respiratory tract (Hannant, unpublished data) and more closely mimic the

"natural" route of infection. Aerosolization of virus avoids concentrating the inoculum at the

sample site as well. Depending upon the strain and titer, clinical signs resulting from

experimental infection can range from mild to severe (Mumford, Hannant et al., 1990). During

natural infection, it is speculated that fewer virus particles are needed to disseminate the disease.

It is rare to isolate a titer of more than 103 EID0/mL from horses naturally infected with EIV

(Mumford, Jessett et al., 1994b). Furthermore, while nebulisation with as little as 102 EIDs0/mL

is capable of causing an infection, a dose of 106 EIDs/mL is required to cause severe clinical

disease (Mumford, Hannant, & Jessett, 1990). In the present study, a variation in previous

techniques was investigated. Infection of individual horses was performed by aerosolization of


Local antibody production is a significant component of protection against infection with

influenza. Various other methods to collect nasal secretions include nasal washes (Boyce, Gruber,

Sannells, &, 2000;Tamura, Funato, Hirabayashi, Kikuta, Suzuki, Nagamine, Aizawa,

Nakagawa, & Kurata, 1990;Tamura, Funato, Hirabayashi, Suzuki, Nagamine, Aizawa, & Kurata,

1991) or in the case of horses, the insertion of tampons into the nasal passage for 15 min (Lunn,

Soboll, Schram, Quass, McGregor, Drape, Macklin, McCabe, Swain, & Olsen, 1999;Nelson,

Schram, McGregor, Sheoran, Olsen, & Lunn, 1998). The use of rayon-tipped proctoscopic

swabs to collect nasal secretions from the nostrils of horses was investigated. Further, the amount

of fluid and antibody retained by the swabs was determined prior to their use in animal studies.

Materials and Methods


Belgian and Percheron draught horses of both sex and approximately 12 months-old were

purchased from a farm in Manitoba, Canada. Prior to their purchase, horses were bled and sera

were screened for antibodies to EIV. Horses that were identified as being sero-negative to EIV

(by HI and SRH) were purchased and transported to the UF equine research center in Ocala,

Florida. To avoid the potential of exposure to viral pathogens typically found in stock-yards, the

draught horses were watered and fed in the trailer. An HI and an ELISA were performed on each

animal again prior to enrolling them in the study. All studies were approved by the UF

Institutional Animal Care and Use Committee (IACUC).

Infection Strains of Equine Influenza Virus

The EIV strain A/equine/2/Kentucky/95 was obtained from Dr. Tom Chambers, from the

University of Kentucky at the second egg-passage. The virus was passage in eggs a further time

and frozen at -700 C until used. The virus was tested for influenza by hemagglutination and

titrated in 10-day-old embryonated chick eggs.

Infection Model Validation Study

A validation study was conducted to establish a reliable method to infect horses with EIV

that would result in seroconversion and clinical disease. In this study, clinical disease was

defined as an increase in temperature over 101.90 F (38.7 C), appearance of ocular or nasal

discharge, anorexia, coughing, and lethargy. Two groups of randomly allocated horses, six in

each with two negative controls, were infected with a "high dose" (108 EIDso/animal) or a "low

dose" (105 EIDsd/animal) of A/equine/2/Kentucky/95 EIV. Horses were housed in individual

stalls within Progress Center's large-animal biosafetyy level 2 facility) research barn during the

infection procedure and up to ten days afterward. Individual horses were infected by

aerosolization of 5-mL suspension of virus-infected allantioc fluid diluted in sterile PBS. An

equine AeorMaskTM (Turdell Medical, South London, Ontario, Canada) connected to a Devilbiss

(Ultra-Neb 99 Ultrasonic, Model 009HD) nebulizer was placed over the nose to deliver

aerosolized virus (Figure 3-1). Horses were monitored daily for the appearance of clinical

disease and bled once a week for two weeks after infection. A clinical scoring system was used

to provide a subjective measure of clinical disease (Table 3-1). Clinical score observations were

made by individuals that were un-blinded to the treatment groups.

Table 3-1. Scoring system associated with clinical signs observed in horses following infection
with equine influenza virus


Rectal Temperature in
Degrees F

Nasal Discharge






Assessed Value

:' "1

Figure 3-1. Infection of horses with equine influenza virus within isolation stalls using a nebulizer
connected to an equine AeroMaskTM

Nasal Secretion Sampling

Horses were placed in stocks with their heads restrained allowing minimal movement. A

16-in rayon-tipped proctoscopic swab (Quality Medical Products, Guilton, MN) was inserted 15-

cm-deep into the ventral meatus of each nostril where it remained in place for five min (Figure 3-

2). Care was taken to avoid trauma to the mucosal lining of the airway. Swabs were placed in

the barrel of a sterile 12-cc syringe and held on ice until centrifugation. Nasal secretions were

collected into the bottom of a polypropylene, 12-cc syringe casing by centrifugation at 4000 rpm

for 20 min at 40 C. The samples were transferred into 1.8-mL cryogenic storage vials and placed

in a 200 C freezer until assayed.

Figure 3-2. Placement of rayon-tipped proctoscopic swabs in horses to collect nasal secretions for
IgA analysis. Swabs were inserted approximately 15-cm-deep into the ventral meatus where they
remained for 5 min. Nasal secretions were recovered from the swabs by centrifugation described
in materials and methods.

Absorption of Antibody by Rayon Swabs

To determine the amount of antibody and fluid volume retained by Rayon swabs, a pre-

measured volume containing a known antibody titer was adsorbed onto duplicate swabs and the

eluted fluid was assayed for antibody titer. A test suspension containing a 1:10 dilution of

equine-specific IgGa monoclonal antibody was assayed prior to and after centrifugation through

the swab at 4000 rpm for 20 min at 40 C in a manner identical to serum samples described in the

previous section on ELISA development. The fluid weight and volume were measured before

being absorbed onto the swab and after centrifugation. The amount of antibody retained in rayon

swabs was determined by 1) measuring the OD reading from an ELISA using primary antibody

supernatant at a 1:10 dilution and 2) obtaining the absorbance of the suspension at 280 nm. One

method of quantifying protein is to measure the absorbance of UV irradiation by proteins at 280

nm. An approximation of the concentration of protein in a solution is determined by the

relationship, 1 absorbance unit is equal to I mg/mL (Peterson, 1983).


Validation of Infection of Horses with Equine Influenza Virus

The techniques and the virus strain used to infect horses resulted in clinical disease

(Table 3-2) and seroconversion (Figure 3-3). The mean clinical score for the high-dose group

was 21.5 and 18 for the low-dose group. In this experiment, six of six horses infected with a

"high dose" (10s EID5/animal) and six of six horses infected with a "low dose" (105

EID5O/animal) of EIV-Kentucky-95 shed virus for an average of 7 and 6 days respectively. The

maximum titer of virus shedding was 1065 EIDs/mL and 105 EIDs/mL for high and low

challenge-dose, respectively. The maximum titers of viral shedding occurred two to three days

later in the low-dose than the high-dose. The non-infected controls did not shed virus.

Table 3-1. Mean clinical scores of horses infected with a "high" (EIDo0 108) and "low" (EIDso
105) dose of equine influenza virus.

Group ID Animal # Animal Group
Total Mean

High Dose 44 24 21.5
45 24
46 29
47 8
48 57
50 11
Low Dose 54 23 18
S 4.1
55 16
56 16
57 22
58 19
59 12
Negative Contros 46 0 < 1
49 0
52 2
51 0

--High dose
-6- Control (high doe)
--Low dose
-O- Costrol (low dose)


Days Postinoculation

Figure 3-3. Serconversion of horses infected with a "high" (EIDso 108) and "low" (EIDso 105)
dose of equine influenza virus. (High dose n = 6; Low dose n = 6; Controls n = 2)

t i




Sampling Method Used to Collect Nasal Secretions

Horses tolerated the placement of swabs into each nostril for the 5-min sampling period

without the use of a sedative. The swab-tips were small enough so as not to occlude breathing. A

clear non-viscous nasal discharge was sometimes noticed dripping from the nostril when the

swabs were in place. No hemorrhaging was noticed in association with the placement and

removal of the swabs. From a total of 138 sampling events, the average volume of secretions was

0.27 mL and 0.54 mL per nostril and horse respectively (Figure 3-4).

0.9 -








Mean Volune Per Horse Mean Volume Per Nostll
Nasal Secretions

Figure 3-4. Mean volume of nasal secretions from horses that were collected by inserting rayon-
tipped swabs into the nostrils for 5 minutes. (Sample n = 138) Error bars represent SE of the

Determination of Antibody Resorption by Rayon-Tipped Swabs

Approximately 10% of the fluid (corresponding to volume) was retained by the swabs in

association with the techniques used to process nasal samples (Table 3-2). An ELISA did not

detect a loss in antibody concentration after absorption in swabs. Furthermore, no loss of protein

was detected by measuring the absorbance of the antibody-containing fluid at 280 nm (Table 3-


Table 3-2. Fluid and Antibody Retention of Rayon Swabs

Swab #1 Swab # 2

Fluid wt. Before 1.88 gm Fluid wt. Before 1.94 gm
Fluid wt. After 1.67 gm Fluid wt. After 1.7 gm
Fluid wt. Retained 0.16 gm Fluid wt. Retained 0.23 gm
% Retention 8.5% % Retention 11.8%

Anti-equine IgG Mean ELISA OD Absorbance at 280 nm

Before Absorption 1.041 1.315
After Fluid Recovery 1.155 1.556


Studies performed at the Animal Health Trust in the U.K. typically include administering

nebulized virus into a room, (Mumford, Hannant, & Jessett, 1990). In the present study,

individual horses were infected with EIV using a nebulizer attached to an AeroMaskT. This

method has potential advantages over other methods by reducing the risk of environmental

contamination and spread of virus. Horses tolerated the placement of the mask which allowed

the direct administration of aerosolised virus particles into the upper air ways. The infectious

dose and method of administration resulted in clinical disease similar to that reported by others

(Mumford, Hannant, & Jessett, 1990). It was interesting to note that while both high and low

dose groups shed maximum titers that were similar, the peak occurred two to three days later in

the low-dose group. This may have been partially due to the additional time required for viral

replication in the low-dose group to catch up to that seen in horses receiving the high-dose. Based

on these data, a challenge dose of 10s EIDso/animal of EIV was used to challenge horses in the

animal study (Chapter 4).

Local antibody formation is believed to be critical for protection against EIV. In recent

studies to determine the level of IgA formation in horses following vaccination and infection,

tampons were placed in the nostrils of ponies. However, the tampon method requires the sedation

of horses. An alternative to the tampon method was employed by placing rayon-tipped swabs

15-cm-deep into the ventral meatus for five min. While this procedure clearly caused discomfort,

no horses required sedation. During the 5-min sampling-period, an average volume of 0.54 mL

of nasal secretions was collected from each horse. The ELISA protocol required only 10 pl of

sample per run. Therefore, this method proved to be adequate in obtaining a sufficient volume of

nasal secretions for the study. Care was taken to avoid trauma to the mucosal surface and no

hemorrhaging was associated with this procedure. This method eliminated procedures to account

for an unknown dilution factor associated with lavage techniques. The alternative use of wicks or

swabs over nasal washes was reported to have higher IgA yields in humans (Thompson, Pham et

al., 1996). In the current study, proctoscopic swabs did not retain a significant volume of fluid

after centrifugation. Furthermore, the amount of antibody lost due to adherence was below the

ELISA detection limits. The method to determine antibody loss included anti-equine IgGa

diluted in PBS. This did not account for potential variation associated with the mucous

component of nasal secretions, however. Based on the results of this study, the use of rayon-

tipped proctoscopic swabs to absorb mucosal fluid from horses was concluded to have advantages

over nasal-lavage and as effective as the tampon method without the sedation of horses.




Equine influenza virus is known to cause upper respiratory disease in susceptible horses

despite their vaccination history. This is, in part, due to the inability of conventional inactivated

vaccines to induce all the components of the immune response. Further, due to the antigenic

changes that occur in influenza viruses and the mobility of the equine community, horses that

receive routine immunizations are nevertheless likely to encounter a variant field strain to that

included in the vaccine. Collectively these events are potentially responsible for maintaining the

endemic nature of this disease. The importance of serum and nasal antibody formation in horses

to prevent infection or reduce the severity of clinical disease has been well established (Mumford

& Wood, 1992) (Lunn, Soboll, Schram, Quass, McGregor, Drape, Macklin, McCabe, Swain, &

Olsen, 1999). In addition, others have described the importance of the cell-mediated arm of the

immune response during infection with EIV(Hannant & Mumford, 1989). Others have reported

the antigen-specific proliferation and cytotoxic T-lymphocyte response induced from either

vaccination or infection with EIV (Hannant & Mumford, 1989;Hannant, 1994). Several reports

have characterized serum or local antibody increases in horses after vaccination with various

adjuvanted inactivated viruses (Wood, Mumford, Folkers, Scott, & Schild, 1983;Mumford,

Wood, Folkers, & Schild, 1988;Mumford, Jessett, Dunleavy, Wood, Hannant, Sundquist, &

Cook, 1994a;Mumford, Wilson, Hannant, & Jessett, 1994;Mumford, Jessett, Rollinson, Hannant,

& Draper, 1994b), modified-live virus (Holmes, Lamb et al., 1991;Lunn, Hussey, Sebing,

Rushlow, Radecki, Whitaker-Dowling, Youngner, Chambers, Holland, Jr., & Horohov, 2001),

and DNA vaccines (Lunn, Soboll, Schram, Quass, McGregor, Drape, Macklin, McCabe, Swain,

& Olsen, 1999;Nelson, Schram, McGregor, Sheoran, Olsen, & Lunn, 1998). However, no

studies describing cytokine mRNA expression in horses following vaccination against or

infection with EIV have been conducted.

In the present study, a comprehensive approach is described to characterize the immune

response, including systemic and local antibody formation, cell mediated immunity, and the

induction of cytokines following immunization with a novel recombinant DNA vaccine. In

addition, a series of identical experiments using a modified-live and inactivated-virus vaccine

were conducted as a comparison. Each vaccine was further assessed for its ability to prevent

infection or severe clinical disease when compared to that seen by natural infection.

Materials and Methods


Sixteen Belgian and Percheron draught horses of both sex and approximately 12 months-

old were purchased from a farm in Manitoba, Canada. Prior to their purchase, horses were bled

and sera were screened for antibodies to either EIV or EHV-1 and 4. Horses that were identified

as being sero-negative to EIV (by HI and SRH) and EHV (by complement fixation assay) were

purchased and transported to the UF equine research center in Ocala, Florida. To avoid the

potential of exposure to viral pathogens typically found in stock-yards, the draught horses were

watered and fed in the trailer. An HI and an ELISA were performed on each animal again prior

to enrolling them in the study. All studies were approved by the UF Institutional Animal Care

and Use Committee (IACUC).


Three vaccines were used in this study. The recombinant DNA vectored vaccine

(Hoescht Roussel Vet, Marburg, Germany) was an experimental product being tested in our

laboratory for efficacy against infection with EIV. The EHV-4 (Dutta strain) "back-bone" vector

(Construct S-4EHV-045), lot # UF-001 consisted of mutations in a large region of 3 genes:

Thymidine kinase = 634 bp, unique short region 2 (US2) = 705 bp, and the glycoprotein E (gE) =

1696 bp (Figure 4-1). Hemagglutinin and NA genes were then inserted with the appropriate

promoter into the portion of the gE gene deletion. Western blot analysis conducted in a

collaborating laboratory (David Bloom, University of Florida) demonstrated HA and NA protein

expression in Vero cell cultures (data not shown). The virus suspension of lysed Vero cells was

determined to contain an infection titer of 106.2 PFU/mL. Two mL of the recombinant vaccine

was administered either by intranasal (IN) or intramuscular (IM) route to two different groups of

horses (n=3). Intramuscular inoculation consisted of administering 2 mL of vaccine virus at the

level of the fifth cervical vertebra, dorsal to the upper edge of the brachiocephalic muscle and

ventral to the fundicular part of the nuchal ligament. Intranasal inoculations consisted of

instilling 2 mL of the vaccine suspension using a size # 8-french polypropylene catheter

(Sherwood Medical, St Louis, MO) inserted approximately 15 cm deep into the left nostril. The

horses were individually housed in an isolation stall equipped with negative airflow and HEPA

filtration in a BSL-2 facility.

Fluvac Plus (Fort Dodge Laboratories, Inc., Fort Dodge IA) is an inactivated-virus

vaccine prepared with types A, (Prague-56) and A2 (Kentucky-92) killed virus in combination

with MetaStim as an adjuvant. This vaccine is reported by the manufacturer to stimulate have

significant neutralizing antibody increases against Alaska-91, Kentucky-91, Kentucky-93,

Kentucky-94, Saskatoon-90, Kentucky-92, Kentucky-95 (type A2 viruses), New-Market-93,

New-Market2-93, Sussex-89, Arundel-91, and Prague-56 (type A, viruses). One mL of vaccine

was delivered by intramuscular route in the same location describe for the recombinant vaccine.

Horses receiving this vaccine were held in a pasture at the UF Equine Research Farm in Ocala,

Florida until challenge-infection.

Flu AvertTM (Heska, Fort Collins, CO) is a modified-live intranasal vaccine containing a

cold-adapted strain of Kentucky-91 (type A2 virus). This vaccine is reported by the manufacturer

to protect against infection with Kentucky-91, Kentucky-98 (American A2) strains and

Saskatoon-90 (Eurasian A2 strain). The vaccine is supplied as a desiccated virus that requires re-

constitution with sterile water. In compliance with the manufacturer's recommendations, the

vaccine was reconstituted less than ten min prior to a 15-cm-deep intranasal inoculation. Animals

receiving this vaccine were held in a pasture at the UF, Equine Research Farm in Ocala, Florida

until challenge-infection. This vaccine was determined in our laboratory to have an HA titer of



Kb0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

-m -

0 /

/ 4-





Figure 4-1. Schematic representation of the recombinant DNA vaccine EHV-4 "backbone" vector
showing gene deletion and insertion points. The genes for the EIV proteins, HA and NA, were
inserted at the gE deletion site.


The EIV strain A/equine/2/Kentucky/95 was obtained from Dr. Tom Chambers

(University of Kentucky) at the second egg-passage. The virus was passage in eggs a further

time and frozen at -700 C until used. The virus was tested for influenza by hemagglutination and

titrated in ten-day-old embryonated chick eggs. The A/equine/2/Kentucky/91 strain contained in

the modified-live vaccine was passage in ten-day old embryonated chick eggs prior to its use in

proliferation assays.

Sample Collection and Storage


Whole blood was collected by jugular venipuncture into 7-mL capacity red-top, serum-

separator tubes (Fisher Scientific, Atlanta, GA) using a 20-g needle. Blood was allowed to

separate for 1 hr at room temperature prior to centrifugation at 2500 rpm for ten min at 40 C.

Sera were stored in four, 1.8-mL cryogenic storage tubes (Nunc, Nenmark) at -20 C in two

separate freezers until assayed.

Peripheral blood mononuclear cells

Whole blood was collected by jugular venipuncture into heparinized green-top tubes.

Peripheral blood mononuclear cells (PBMC) were isolated as previously described (Allen,

Yeargan, Costa, & Cross, 1995) with a FICOLL-Histopaque 1.077 g/mL density gradient (Sigma

Chemical Co., St. Louis, MO). The cells were washed three times in phosphate buffered saline

(PBS) (Gibco BRL, Grand Island, NY) and pelleted by centrifugation at 250 x g for ten min at

room temperature. One mL of red blood cell (RBC) lysing buffer (8.3 g/L ammonium chloride in

0.01 M Tris-HCL buffer) (Sigma Chemical Co.) was added to the cell pellet if RBC were present.

The PBMC were washed an additional two times if RBC lysis buffer was used. For long-term

storage, PBMC were suspended in a freezing medium consisting of RPMI-1640 with 10-mM

Hepes buffer, 2-mM L-glutamine, .075% w/v sodium bicarbonate, 1 mM sodium pyruvate, 100

U/mL penicillin-G, 100 ug/mL streptomycin, 10% BFS, and 50% bovine fetal serum (BFS)

(Gibco BRL, Grand Island, NY) at a density of approximately 1 x 107 cells/mL. The PBMC were

placed in a NalgeneT" Cryo 10 C freezing canister (5100-001) containing 70% isopropyl alcohol.

The container was placed in a 800 C freezer over-night to achieve a controlled rate of freezing.

The tubes containing frozen cells were subsequently transferred to liquid nitrogen (cells were

stored in the liquid phase).

Nasal secretions for IgA antibody determination

Horses were sampled by methods described in chapter three under standardization and

development of animal infection and sampling methods.

Experimental and challenge infection

Horses assigned to the EIV-infection group received an initial infection at an infectious

dose of 106.5 EIDs0 on post infection day (PID) zero. Horses were housed in individual stalls

within Progress Center's large-animal biosafetyy level 2 facility) research barn during the

infection procedure and up to ten days afterward. Individual horses within the vaccine and EIV-

infection groups were challenged with an infectious dose of 10S EIDs/animal by methods

described in chapter 3.

Virus isolation

A 16-in rayon-tipped proctoscopic swab was inserted approximately 15-cm-deep into the

ventral meatus of each nostril where it remained for 15 sec. The swabs were placed into

polystyrene tubes containing viral transport medium (VTM). The VTM was composed of

Dulbecco's Modified Eagle Medium with 25 mM Hepes, 5% BFS, penicillin-G sodium (100