Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Comparative analysis of rubella specific antibody responses in congenitally and postnatally rubella infected… Mauracher, Christoph Andreas 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_fall_mauracher_christoph_andreas.pdf [ 3.68MB ]
JSON: 831-1.0098879.json
JSON-LD: 831-1.0098879-ld.json
RDF/XML (Pretty): 831-1.0098879-rdf.xml
RDF/JSON: 831-1.0098879-rdf.json
Turtle: 831-1.0098879-turtle.txt
N-Triples: 831-1.0098879-rdf-ntriples.txt
Original Record: 831-1.0098879-source.json
Full Text

Full Text

(Q’\  COMPARATIVE ANALYSIS OF RUBELLA SPECIFIC ANTIBODY RESPONSES IN CONGENITALLY AND POSTNATALLY RUBELLA INFECTED HUMANS: A MODEL FOR SELECTIVE TOLERANCE  by  Christoph Andreas Mauracher B.Sc., The University of Victoria, 1986  A thesis submitted in partial fulfilment of  the requirements for the degree of Doctor of Philosophy in the Faculty of Graduate Studies Pathology Program  We accept this thesis as conforming to the required standard  The University of British Columbia 1992 © Chris A. Mauracher  In presenting this thesis in  partial fulfilment of the requirements for an advanced  degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department. or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  -  /  The University of British Columbia Vancouver, Canada Date  ?y 6’  DE-6 (2/88)  D7 I  ABSTRACT Rubella virus (RV) is an enveloped, positive sense, single stranded RNA virus in the family Togaviridae. The virus consists of a host derived lipid bilayer membrane, two trans-membrane proteins El and E2, and the C protein which together with viral RNA forms the icosahedral nucleocapsid.  RV causes natural infections in  humans only, where it can lead to a variety of pathological conditions. The most severe outcome, Congenital Rubella Syndrome (CRS), occurs when the fetus is infected in the first trimester of gestation.  RV can be  isolated from CRS patients at birth and is believed to establish a persistent infection. Although congenital RV infection remains of clinical importance it also provides a rare opportunity for studying the effects of intrauterine viral infection on the human fetus. This thesis examines the hypothesis that intrauterine infection with RV can establish immunological tolerance which is reflected in the serological response to RV. The study on RV-specific tolerance may help in the understanding by which mechanism self-recognition is established in the human. The induction of immunological tolerance may also be an important mechanism in viral persistence and play a role in chronic inflammatory disorders such as Rubella Associated Arthritis (RAA). A comprehensive study of the antibody mediated immunity of healthy individuals responding to postnatal infection with RV is compared to that of CRS patients. Assays were developed to measure antibody quantity, affinity and kinetics of the antibody response to whole virus as well as that to the individual RV structural proteins, El, E2 and C.  The viral proteins were purified from whole RV preparation by differential  centrifugation followed by preparative SDS-PAGE under non-reducing conditions.  Separated proteins were  analyzed for their structural integrity by assaying their biological activity. RV protein-specific ELISAs were developed and used for antibody quantitation and in IgG affinity assays using the chaotropic elution technique. Biological activity of sera was assayed by HAl assay. The specificity of antibodies to linear and topographic epitopes was investigated by using reducing and non-reducing Western blotting. Observations made during the development of protein separation protocols have led to the description of an uncoating mechanism for RV. While establishing ELJSA protocols, improved buffers were developed by including heat denatured blocking proteins and studies on a novel anti-RV IgG ELISA, which will give a more accurate assessment of protective immunity than ELISA technology used to date, were initiated. Results indicate that CRS patients can produce similar amounts of IgG than control patients if measured by whole RV ELISA. Group differences were detected at the level of the protein specific responses. CRS patients exhibited significantly lower levels and had reduced affinities of El-specific antibodies.  The most consistent  feature of this patient population was their inability to produce IgG to linear epitopes of the El protein. The potential role of the tolerization of anti-El Th cells due to the exposure of the immature fetal immune system to RV is discussed.  A model is proposed which can accommodate  the fmdings on serological and  lymphoproliferative immune responses of CRS patients and which supports the hypothesis that congenital RV infection in the early gestational period leads to viral antigen tolerization. Tolerance to the immunodominant El protein may lead to a sufficient depression in responder T and B cells in order to allow chronic, low grade RV replication. II  TABLE OF CONTENTS ABSTRACT TABLE OF CONTENTS  iii  LIST OF TABLES  vii  LIST OF FIGURES  viii  LIST OF ABBREVIATIONS  ix  ACKNOWLEDGEMENTS  xi  1. INTRODUCTION  1  1.1. Rubella Virus Bioloay  1  1.1.lClassification  1  1.1.2RV Morphology and Physio-Chemical Properties  2  1.1.3 Structural Proteins  3 El Glycoprotein  3  l.1.3.2E2 Glycoprotein  4 Protein  5  1.L4.Non-structural Proteins  6  1. 1.5.Genome Organization  6  1.1.6.RV Life Cycle  7 Entry and Uncoating  7  1.1.6 .2 Replication and Protein Expression  8  9 10  1.2. Rubella Pathology and Pathoenesis 1.2.1.History  10  1.2.2.Clinical and Immunological Features  10  l.2.2.lRubella  10  1 .2.2.2.Congenital Rubella Syndrome  13  a) Clinical Features  13  b) Immunological Features  17 .Tolerance following Intrauterine Infection 1.3. Rubella Vaccine  18 21  1.3.1.History  21  1.3.2.Biology and Immunology  22  1.3.3.Side Effects of RV Vaccination  23  1.3.4.Vaccine Efficacy  25  1.4. Laboratory Diaanosis of Rubella Virus Infection  25  1.4.1.Virus Isolation  25  1.4.2.Neutralization Assay  26 111  l.4.3.Hemagglutination Inhibition Assay  27  1.4.4.Enzyme Linked linmuno-Assays (ETA)  27  28  1 .4.4.2.Western Blotting  29 29  1.4.4.Affinity Assays .Chaotropic Elution ELISA  32 EUSA  33  1.5. Thesis Rationale and Objectives  33  2. MATERIALS AND METHODS  36  2.1. Virus Preparation. Titration and Concentration  36  2.l.1.BuIk Preparation of Rubella Virus (M-33 Strain)  36  2.l.2.Rubella Virus Titration on RK-13 Cells  36  2.1.3.Concentration of Rubella Virus Tissue Culture Supernatant  36 37  2.2. Preparative SDS-PAGE 2.2. 1.Electrophoresis  37  2.2.2.Electroelution  38  2.2.3.Detergent Extraction  38  2.2.4.Yield Determination  38  2.3. Metabolic Radiolabelling of Rubella Virus  39  2.4. Serum Separation  39  2.5. Gel Stainin2  40  2.5. 1.Colloidal Coomassie Stain  40  2.5.2 Silver Stain  40 40  2.6. Solid Phase linmunoassays 2.6.1.Western Blotting  40  2.6.2.Iinmunoprecipitation  41  2.6.3.ELISA  42 RV ELISA  42 Protein ELISA  43 43  2.7. Affinity Assays 2.7. 1.Chaotropic Elution ELISA  43  2.7.2.One Well Inhibition ELISA  44  2.8. Hemaalutination Inhibition Assay  45  2.9. Biological Function of Rubella Virus Proteins  45  2.9. 1.Solubility Shift of C Protein under Acidic pH  45 .Ultracentrifugation  46  2.9.1 .2Polymerase Chain Reaction for the Detection of RV RNA  46  2.9.2.Hemagglutination Activity of Purified El Protein iv  47  3.  2.1O.Patient Sera  47  2.11. Statistical Methods  48  RESULTS AN]) DISCUSSION  49  3.1. Section I: Virus Preparation and Use in Solid Phase Iminunoassays  49 49  3.L1.Rubella Virus Preparation  49 Titration  51  3.1.2.Whole RVELISA Conditions and Antigen Treatment  51 Denatured Sample Buffer  52 54  3. 1.3.Afflnity Assays Well Inhibition Assay  55 Urea Elution ELISA  57 of IgG Fractions from Human Sera  58  3.2. Section II: RV Protein Specific Immunoassays  62 62  3.2.1.Western Blottina 3.2.1. lReducing vs. Non-Reducing Conditions  62 of Immunoprecipitation and Western Blot  64 Blot AffinityAssay  66  3.2.2.Separation of RV Proteins, Their Use in Solid Phase Immunoassav  67  3.2.2.LFeasibility of Electroelution; Yield and Purity of RV Proteins  69 of Protein Binding Conditions  71 of Anti-El, -E2 and -C IgG by ELISA  71  3.2.3.Antiaenicity and Bioloaical Activity of Separated RV Structural Proteins  72 of El  74 of C  75 81  3.2.4.Summarv  3.3. Section ifi: RV-Specific Humoral Immune Response in Adults Following Rubella  82  3.3.1.Response to Whole Virus and Separated RV Proteins  82  3.3.2.Biological Activity of Sera  82  3.3.3.Reactivitv to Linear and Topographic Epitopes  83  3.3.4.Kinetics of the laG Response  84  3.3.5.Relative Affinity of IG Directed to RY and RV Proteins  86  3.3.6.Summary and Conclusion  88 Responses to RV Proteins  89 Protein Specific LymphoproliferativeResponses  90  V  3.4. Section IV: RV-Specific Humoral Immunity in CRS Patients  92  3.4.1.RV and RV Protein Specific ELISA  92  3.4.2.Biolopical Activity of Sera in CRS Patients  94  3.4.3.Relative Affinity of RV Specific JaG in CRS Patients  95  3.4.4.Reactivity to Linear and Topoaraphic Epitopes in CRS Patients  96  3.4.5.Analysis of Sequential Serum Samples from CRS Patients  99  3.4.6.Evidence for Th Tolerance in CRS Patients  101  3.4.7.Summary and Discussion  104  4.  SUMI’VIARY AIM]) PERSPECTIVES  108  5.  REFERENCES  112  vi  LIST OF TABLES  Table Table Table Table  1: 2: 3: 4:  Table 5: Table 6: Table 7: Table 8: Table 9:  Summary of Physical and Immunological Properties of RV Structural Proteins. Clinical Features of Congenital Rubella. Mechanisms of Viral Persistence. Biological Reactions in Which High Affmity Antibodies are Superior to Low Affinity Antibodies. Rubella IgG ELJSA Responses to Whole RV and RV Structural Proteins. Relative Affmity of Anti-RV IgG in Serum or Purified IgG. Percentage of Signal Remaining on Western Blot under Reducing Conditions. Percentage of Signal Remaining on Western Blot under Reducing for Control and CRS Patients Serological Examination of MMR Vaccinated CRS Patients.  vii  6 15 19 31 54 60 84 97 101  LIST OF FIGURES  Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure  1: Model of Rubella Virus Structure. 2: Model of the E1/E2 Glycoprotein Spike of RV. 3: Strategy for the Expression of RV Structural Proteins. 4: Model for the Topogenesis of the RV Polyprotein. 5: Clinical Features of Rubella. 6: Incidence of Virus Shedding in CRS Patients. 7: Incidence of Reported Rubella and CRS Cases in the USA. 8: Hemagglutination-Inhibition (HAl) Test. 9: Triton X-1 14 Extraction of RV from Tissue Culture Supernatant. 10: Micrograph of RV Infectious Focus on RK-13 Cell Monolayer. 11: Western Blot and Total Protein Stains of RV Preparations Separated on SDS-PAGE Gels. 12: Titration of RV Treated with Triton X-100. 13: Determination of RV-Specific IgG by ELISA. 14: Determination of 10.5 in Acute and Convalescent Sera. 15: Affinity Maturation of Anti-RV IgG Measured by Inhibition ELISA. 16: 8 M Urea Elution ELISA for Determination of Antibody Affinity. 17: Treatment of Solid Phase Bound RV with 8M Urea. 18: FPLC Separation of Human Serum. 19: Western Blot and Autoradiograph of Reduced and Non-Reduced RV Preparations. 20: Comparison of RV IgG Western Blot and Immunoprecipitation Assays. 21: Immunoblot Analysis of RV-Specific IgG, 1gM and IgA. 22: 8M Urea Elution of Western Blot (North-Western Blot). 23: Yield of Radiolabelled RV Proteins after Electroelution. 24: Western Blot of Separated RV Proteins. 25: Comparison of Optimal Coating Concentrations for El, E2 and Cd. 26: Western Blot and Densitometric Scan of Anti-RV Standard Serum. 27: Effect of SDS and Temperature on the Antigenicity of Whole RV. 28: HAl Antibody Inhibition Assay. 29: Autoradiographs of TX-i 14 Extracted RV. 30: Ultracentrifugation of RV Following Detergent Extraction. 31: Proposed Model for the Entry of RV into the Host Cell. 32: Detergent Extraction of Separated C Protein. 33: Quantitation of IgG to Whole RV and RV Proteins. 34: The Kinetics of the IgG Response to RV Structural Proteins. 35: Differential IgG Affinity Maturation to the Structural Proteins of RV. 36: Kinetics of the Appearance of RV Protein-Specific IgA. 37: Comparison of Proliferative and IgG Responses to RV Structural Proteins. 38: RV-, and RV Protein-Specific IgG Responses in Control and CRS Patients. 39: Affmity Indices of IgG Directed to RV and RV Proteins in Control and CRS Patients. 40: Western Blot under Reducing and Non-Reducing Conditions. 41: Model for the Production of Anti-El IgG in the Absence of El-Specific Th Cells. 42: Western Blot Analysis for Three MMR Vaccinated CRS Patients. viii  2 4 8 9 12 16 22 28 46 49 50 52 53 55 56 57 58 59 63 64 65 66 68 69 70 72 73 75 76 78 79 80 83 85 87 89 90 93 95 97 98 100  LIST OF ABBREVIATIONS 405 A  Absorbance at 405 nm wavelength  aa  amino acid  AU  Arbitrary Units  BCIP  5-Bromo-4-Chloro-3-Indolyl Phosphate  BDV  Border disease virus  Bis  N,N’ -methylene-bis-acrylaniide  bp  base pairs  BSA  Bovine serum albumin Degrees Celsius  cb  Conjugate Buffer  CIC  Circulating Immune Complex  CRBC  Chick Red Blood Cells  CRS  Congenital Rubella Syndrome  EIA  Enzyme linked inimuno assay  ELJSA  Enzyme linked immunosorbent assay  ER  Endoplasmic reticulum  et al.  et alii  HA  Hemagglutination  HAl  Hemagglutination Inhibition  Ig  Immunoglobulin  IU  International Units  kDa  kilodalton  LCMV  Lymphocytic Choriomeningitis Virus  M  molar  MBq  Megabecquerel  MEM  Minimum Essential Medium  nil  millilitre  11  microlitre  im  micrometer  MMR  Mumps, Measles and Rubella vaccine  MW  Molecular Weight  n  sample number  NBT  Nitro Blue Tetrazolium  NGS  Normal goat serum  nm  nanometer  ns  non-structural  NT  Neutralization titre  PAGE  Polyacrylamide Gel Electrophoresis  PBS  Phosphate buffered saline  pfu  Plaque forming units  pH  1/log [Hj ix  p1  isoelectric point  rpm  Rotation per minute  RA  Rheumatoid Arthritis  RAA  Rubella Associated Arthritis  RT  Room temperature  RV  Rubella virus  S  Svedberg unit  sb  Sample Buffer  sbb SD  Sample Buffer (boiled) Standard Deviation  SDS  Sodium Dodecyl Sulphate  SFV  Semliki Forest virus  SI  Stimulation Index  subb  Substrate Buffer  SV  Sindbis virus  Tc  Cytotoxic T cell  Th  Helper T cell  TCS  Tissue culture supernatant  Tris  trishydroxymethylaminomethane  TX-100  Triton X-lOO  TX-114  Triton X-114  VSV  Vesicular Stomatitis virus  wb  Wash Buffer  x  ACKNOWLEDGEMENTS I want to thank the members, past and present, of the lab: Fiona Yong, Dr. Liz Hancock, Susan Farmer, Margaret Ho, Diane Ddcarie, Ting Zhang, Dr. Beatriz Gomez, Janet Deronio and Dr. Marita Lundström Hobman.  Special thanks go to Bob Shukin who in most cases was a thorn in my side but whose ideas,  discussions and experience helped in many crucial experiments and to Susan Farmer for the growing of much of virus that I used. To Dr. Aubrey Tingle, my supervisor, I would like to extend my gratitude for the patience patients  -  -  not to mention his  he had in taking me on as a graduate student and for defending me where others may have given up.  If I emerge as a Doctor of Philosophy, it is his coaching which will have made it possible. My co-supervisor, Dr. Leslie Mitchell, I would like to thank for creating the stable lab environment which she created in the last years and especially for convincing me that all polyvinyl surfaces are not the same. Thank you to my supervisory committee, Drs. O’Kusky, Pritchard, McMaster, Chantler and Gillam, for their investment of time and their wisdom to accept this thesis. I appreciated Dr. Shirley Gillam’s help and advice on my questions regarding anything to do with Rubella virus and for the many calories which were consumed on her account. Of the agencies and individuals who funded this study I would like to thank the Research Division of the Children’s Hospital of British Columbia, University of British Columbia and especially Mr. Roman Babicki who through a personal gift of fmancial support and enthusiasm enabled much of this work. It is easy for the scientist to forget that it is only through the interest and sacrifice of the community that research at university institutions is possible. Mr. Babicki personifies this attitude which in many ways is the root for progress in science. Foremost I would like to thank my family. To my parents who endured the tortuous path of digress and success and were there with both moral and fmancial support and to my grandmother, Berta Mauracher, who was behind me since I can remember  -  no matter what. To my wife, Aileen, who put every thing into perspective, failures,  progress and life in general. To the life she gave me, my two beautiful daughters Andrea and Marita, I dedicate this thesis. By the time both of you chance to read these pages much time willhave elapsed, and ifI should have forgotten, remind me of what I promised at your births.  xi  Much in research is doing the wrong thing at the right time, science begins when one fmds the courage to admit it.  xli  1. INTRODUCTION  Rubella virus (RV) causes a mild childhood illness but can cause severe birth defects if the fetus is infected in the first trimester of gestation. The teratogenic effect of RV can cause a wide range of birth defects and late onset sequelae collectively described as the Congenital Rubella Syndrome (CRS). RV persists in utero and can be isolated from virtually all CRS patients at birth and is believed to cause a low grade persistent infection in most patients over many years. The immunological dysfunction which allows the virus to persist and replicate in these individuals has not been defined. The investigations whose results are presented in this thesis examined the humoral responses of adults following the uncomplicated resolution of RV infection in comparison to the RV-specific humoral response of CRS patients to test the hypothesis that intrauterine exposure to a viral agent leads to virus specific immunological tolerance and that specific tolerance is reflected in the serological response to RV. Tolerance or immunological non-responsiveness leading to viral persistence is not only of interest to the pathology of CRS but may play an important role in the pathogenesis of other RV-associated syndromes such as Rubella Associated Arthritis or Chronic Progressive Rubella Panencephalitis, by the chronic low grade replication of virus in specific tissues such as the synovium or the brain. This introduction contains three sections: a summary of rubella virus biology, rubella virus induced pathology and immunity, and a review of the use and development of diagnostic tools in the detection of RV and the RV specific immune-responses.  The rationale and objectives of this thesis will conclude the chapter.  1.1 Rubella Virus Biolo2y 1.1. 1.Classification Rubella Virus (RV) is the only member of the genus Rubivirus in the family Togaviridae (Porterfield et al., 1978). The Togaviridae are defined as spherical, enveloped viruses of 60-70 mu diameter, with an icosahedral nucleocapsid.  The genome is composed of a 40S (+)-sense-single stranded RNA molecule.  Studies on the  replication cycle of these viruses showed significant differences between flaviviruses and the other genera of the Togaviridae, resulting in the reclassification of the flaviviruses into their own family Flaviviridae in 1985  —1—  (Westaway et al., 1985).  The Togaviridae are now composed of the rubivirus, aiphavirus, pestivirus and  arterivirus genera. The old classification of Group A and B Arboviruses is no longer officially recognized by the International Committee on Toxonomy and Nomenclature of Viruses.  1.1.2.RV Morphology and Physiochemical Properties Early electron microscopic studies of RV have shown the virus to be spherical and of 60-70 nm in diameter (von Bonsdorff and Vaheri, 1969). An electron dense core of 30 nm diameter defines the nucleocapsid. An electron lucent ring separating the capsid from the membrane is characteristic and allows this virus to be distinguished from other morphologically similar togaviruses (Murphy et al., 1968). Spikes of 5-8 nm length are observed on the virion surface which are associated with hemagglutination activity.  El E2 C Membrane  Figure 1. Model of Rubella Virus Structure. Capsid protein (33 kD) forms the icosahedral capsid surrounding the viral RNA (10 kb). The nucleocapsid is enveloped by a host derived lipid-bilayer membrane in which envelope proteins El (58 Kd) and E2 (42-47 kD) are embedded. The envelope proteins form El-El homodimers and E2-El heterodimers to form surface spikes of 6-8 urn in length.  -2-  The virus is heat labile and can be rapidly inactivated by exposure to 56 1C. The virus remains stable for several days at 4#C and can be stored for indefmite periods at -70fC. Agents which extract lipids, denature proteins or interfere with nucleic acids all lead to virus inactivation (Parkman et. al, 1964). RV is assembled from three structural proteins: El, E2 and C (Vaheri and Hovi, 1972; Payment et al., 1975). The envelope proteins El and E2 are both type 1 trans membrane proteins, acylated and highly glycosylated. In the intact virion these proteins form El-E2 heterodimers and El-El homodimers, visible by electronmicroscopy. The capsid is composed of C protein and has an icosahedral symmetry of T= 3. This icosahedron surrounds the positive sense, single strand RNA to form the nucleocapsid.  1. 1.3.Structural Proteins l.l.3.1.El Glycoprotein The El glycoprotein is the largest of three RV structural proteins with an apparent molecular mass of 58 kD on reducing SDS-PAGE and an acidic p1 of 6.5 (Ho-Terry and Cohen, 1982).  The protein is a type 1  transmembrane protein, containing a 14 residue cytoplasmic tail and a 27 residue trans membrane spanning region (Vidgren et a!., 1987). The remaining 440 residues form the bulk of this protein’s surface region. Post translational modifications include both fatty acid acylation in the C-terminal region (Waxham and Wolinsky, 1985a) as well as N-linked glycosylation at all three potential glycosylation sites, predicted by sequence analysis of all RV strains studied (Frey et al. 1986; Clarke et al., 1987; Terry et al., 1988). The glycosylation with these endo-H resistant glycans adds 5 kD to the molecular mass of this protein and although no direct biological function for this modification has been described, it is believed that it stabilizes El in its biologically functional and immunologically reactive conformation (Ho-Terry et al., 1984). El contains the hemagglutinin activity as well as the majority of the defined viral neutralization domains. The hemagglutination of day-old chick erythrocytes and pigeon erythrocytes is a function of the El monomer (HoTerry and Cohen, 1985) but does not correlate completely with neutralization domains (Green and Dorsett, 1986). By using trypsin and S.aureus cleavage product analysis the HA activity was localized to a 13 kD fragment of El. This region, containing amino acid residues 245-285,encodes three separate epitopes conferring both neutralizing and HAl antibody binding sites (Terry et al., 1988), showing that these two functions are  -3-  closely situated on the El protein. Using competitive inhibition assays with a panel of monoclonal antibodies it was further shown that at least six non-overlapping epitopes exist on this protein (See summary in table 1). Because of the concentration of 6 of the known 7 neutralizing epitopes and the receptor-like activity of HA, it is believed that El contains the as of yet undefined viral attachment site.  El  E2  Figure 2. Model of the El/E2 Glycoprotein Spike of RV. N-linked sugars are indicated by (Mf located on both El and E2. 0-linked sugars (.f.) are located on E2 only. Fatty acylation ( ) on cysteine residues are possible on both El and E2 cytoplasmic domains.  1.1 .3.2.E2 Glycoprotein E2 is the smaller of the two envelope proteins and travels as a diffuse band in electrophoresis with an apparent molecular mass of 42-48 kD, under reducing conditions. In Therien strain and to a lesser extent in M-33 this diffuse band can be resolved into species termed E2a and E2b (Oker-Bloom et al., 1983). The protein has a wide range of isoelectric points ranging from pH 5.0 to 8.6 (Ho-Terry and Cohen, 1982) which is believed to be due to heterogeneous glycosylation with terminal sialic acids. The protein is highly glycosylated and if Vero  -4-  cells are infected with RV, in the presence of tunicamycin, a 30 kD species of E2 is isolated (Oker-Blom et a!., 1983), indicating that more than one third of this protein’s mass is derived from oligosaccharide chains. Depending on the RV strain, E2 contains 3 (M 33 and HPV 77) to 4 (Therien and RA 27/3) N-linked glycosylation sites (Clarke et al., 1987; Vidgren et al., 1987; Frey and Marr, 1988; Nakhasi et al., 1989) which are thought to be of the complex type, as judged by their resistance to endo-H glycosidase (Oker-Blom et al., 1983). Recently the presence of sialylated 0-linked sugars on E2 has been described (Lundstr6m et al., 1991), clarifying the existence of the 15 or more isoelectric variants of this protein. No direct biological function has been assigned to this protein although the co-expression of E2 with El is necessary for the transport of El into the assembling virion (Hobman et al., 1990). Also one neutralizing domain has been described for this protein (Green and Dorsett, 1986). It has to be borne in mind that E2 occurs in a heterodimeric form with El and in that function might confer support or protection to the spike structure.  1.l.3.3.C Protein Capsid is a non-glycosylated protein of apparent molecular mass of 33 ‘000 under reducing SDS-PAGE.  On  these gels, C is often observed to run as a doublet, differing in less than 1 kD in mass. This is thought to be due to the selective usage of two translation initiation sites separated by 7 amino acid residues (Clarke et al., 1987). Under non-reducing conditions, C runs in a dimeric form with an apparent molecular mass of 66 kD (Mauracher et a!., 199 la). C has two distinct p1 forms at pH 8.8 and 9.5. This protein specifically interacts with the 40 S viral RNA (Weiss et al., 1989) and initiates the assembly of the nucleocapsid.  The exact site and  mechanism of nucleocapsid assembly has remained controversial, as distinct differences in RV are observed in comparison with the well studied alphavirus assembly pathway. Alphavirus capsid contains protease activity and cleaves itself, by an autolytic event, from the El/E2/E3 polypeptide, followed by capsomere assembly in the cytoplasm (Malancon and Garoff, 1987). The capsid of RV has no inherent protease activity (Oker-Blom et al., 1984) and assembly of the capsomers has been observed to be dependent on membranes (Horzinek, 1981) suggesting that RV carries with it the signal peptide of the E2/El polyprotein (Suomalainen et a!., 1990). Four epitopes have been defmed on this protein, none of which induce neutralizing antibodies (Waxham and Wolinsky., 1985a), although capsid has been implicated in the events of viral penetration (Mauracher et al., -5-  1991a).  Protein  MW  Glycosylation  Epitopes  Neutralizing 2 Domains  3 N-linked  6  strong  yes  strong  2  weak  no  weak  4  no  no  medium  El  58’OOO  E2  42-48’OOO  3-4 N-linked/O-linked  C  33’OOO  None  ‘  HA Activity 3  Capacity to Induce T Cells 4  Table 1: Summary of Physical and Immunological Properties of RV Structural Proteins. Epitopes are defmed by competitive inhibition using monoclonal antibodies (Waxham and Wolinsky, 1985b). 2 Defined by monoclonal antibodies which can inhibit viral replication on Vero cells (Waxham and Wolinsky, 1985b; Green and Dorsett, 1986). The ability of RV to agglutinate day old chick- or pigeon erythrocytes has been mapped to El by using purified protein (Ho-Terry and Cohen, 1980) or by monoclonal antibody (Green and Dorsett, 1986). Defined by studying lymphoproliferative responses of healthy adults to RV proteins El, E2 and C expressed individually in vaccinia virus (Chaye et al., 1992).  1. 1.4.Non-Structural Proteins Little is known about the nature of RV non-structural (ns) proteins although their genes occupy two thirds of the RV genome. Most work on togavirus enzymes has been performed in SFV and SV and by analogy with these well studied viruses and findings of RV gene arrangement and sequence, it is believed that RV encodes 4 ns proteins (Schlesinger, 1987). One of these ns proteins is likely functioning as an RNA dependent RNA polymerase. The catalytic steps performed by this protein or protein complex include: initiation and elongation of full length (+) and  (-)  sense RNA, (+) sense sub-genomic RNA synthesis as well as capping and methylation  (Schlesinger and Schlesinger, 1990). Furthermore it is believed that one ns protein acts as a virus specific protease involved in the cleavage of the viral polyprotein.  Host cell proteins may also be involved in the  enzymatic activities during viral replication as prolonged inhibition of host cell RNA synthesis prior to virus infection interferes with togavirus replication (Baric et al. 1983).  1. 1.5.Genome Organization The RV genome exists as an infectious, single-stranded RNA molecule with a sedimentation rate of 40S (Hovi and Vaheri, 1970). In the infected cell, an additional 24S subgenomic RNA can be isolated, similarly poly -6-  adenylated and capped as the full-length genome and with an identical 3’ sequence. This mRNA gives rise to a polyprotein from which the RV structural proteins are derived (Oker-Blom et al., 1984a). The sequence for the 24S subgenomic RNA is known for wild type isolates M-33 and Therien strain (Clarke et al., 1987; Frey et al., 1986) as well as for vaccine strains RA 27/3 and HPV 77 (Nakhasi et al., 1989; Zheng et al.,1989). The aligned sequences of the 24S RNA reveal a 95% homology, whereas little homology exists between the genomes of RV strains and members of the aiphavirus genus (Frey and Marr, 1988).  1.1.6.RV Life Cycle The life cycle of RV is very similar to the replication strategy of the alphaviruses. In particular, the biology of SFV and SV have been studied in much detail and much of the knowledge obtained by the study of these two viruses has been shown to apply to RV as well, although differences in some of the details of morphology and replication exist.  1.1.6.lEntry and Uncoating Togaviruses are believed to enter the target cell by receptor mediated endocytosis, a step which can be selectively inhibited by lysosomotrophic drugs (Helenius et al., 1984). The entry mechanism of the aiphaviruses has been described in detail (Kielian and Helenius, 1986) and there is evidence that RV also uses the endocytic pathway for cell entry (Vaananen and Kaariainen, 1980). A specific cell surface receptor has not been defined for RV, but it has been speculated that membrane phospholipids might be involved in virus attachment (Mastromarino et al., 1990); while others interpret an infection frequency of less than 10% on a cultured cell monolayer as an indication that the RV receptor is a cell cycle dependent protein (Hemphill et al., 1988). Nevertheless, electron microscopic evidence shows RV particles being taken up by endocytosis and delivered to the endosome system (Dr.S.Katow, NIH, Japan, personal communication).  Upon acidification of the  endosome below pH 5.5 the RV El undergoes a structural change to become fusogenic (Katow and Sugiura, 1988), a step which allows the limiting membranes of the RV and endosome to fuse. The low pH environment of the endosome also causes a shift in RV C protein solubility, leading to nucleocapsid uncoating in the endosome (Mauracher et al., 1991). Upon membrane fusion the viral RNA penetrates into the cytoplasm and viral replication can be initiated. -7-  1.1 .6.2 Replication and Protein Expression Naked RV RNA is infectious (Maes et al. 1966). Hence, the first event in replication is the translation of the non-structural proteins needed to produce full length 40S (+)-sense RNA and 24S (+)-sense subgenomic RNA, via a (-)-sense RNA intermediate (Oker-Blom, 1984b). The 24S subgenomic RNA  is  required as a template  for translation of the structural proteins. This message is capped, polyadenylated and from the 5’ end encodes for: C, E2 and El (Figure 3). The molecular mechanism which controls the production of 24S mRNA for structural protein translation, or full length 40S RNA for nucleocapsid packaging remains unclear.  40 S (—11.000 b) AA(A) 1  —  24 S (—3500 b) 5l  AA(A)3 1  p110 ; 2 NH  Ii :  I  .  I  -  J,  capsid —  C 33K  translation  ‘  COOH processing  envelope  —II  E2 30K ,,  II  El 53K glycosylation  E2a (47K) E2b (42K)  ,  58K  Figure 3. Strategy for the expression and processing of RV structural proteins from a 24S + sense RNA sub genomic fragment. (From: Oker-Blom, 1984b)  C protein remains in the cytoplasm whereas the E2 and El proteins are transported to the endoplasmic lumen and processed as transmembrane proteins. Therefore it is believed that translation, proteolytic cleavage and post-translational modifications occur simultaneously at the different sites of this polyprotein. A translocation signal sequence at the COO terminal of the C protein allows the entry of the polyprotein into the lumen of the -  ER where a series of trans-membrane and signal sequences allow for the correct insertion of El and E2 (illustrated in detail in figure 4).  -8-  A  E2  C NH  El I  I  I  I **  **  capsid  B N cytoplasm  ER lumen E2  Figure 4. Model for the Topogenesis of the Rubella Virus Polyprotein. A) The 110 kD polyprotein is shown with the location of signal peptide sequences (white boxes) at the COOH termini of C and E2 and the transmembrane anchor sequences of E2 and El (black boxes). The asterisks (**) indicate published N-terminus sequences of El and E2 (Kalkkinen et a!., 1984). The polyprotein is not drawn to scale. B) Two cleavages by signal peptidase (white arrow head) release the capsid and separate the E2 and El proteins. A third cleavage site (black arrow head) has been proposed to occur in an arginine-rich region in the cytoplasmic tail of £2 (Vidgren et al., 1987). (Figure B adapted from: McDonald, 1991).  The membrane proteins are extensively post-translationally modified. Both El and E2 are acylated and N glycosylated (Waxham and Wolinsky, 1985a; Hobman and Gillam, 1989). E2 is additionally glycosylated with 0-linked saccharides (Lundstr 6m et al., 1991). Although El and E2 are co-translationally inserted into the membrane, their fates in intracellular transporting are quite different. E2 can be fully processed alone whereas El is dependent on E2 for complete processing and for transport to the cell surface (Hobman, 1989).  1. l.6.3.Assembly Depending on the host cell, the RV envelope proteins may either congregate in the plasma membrane or in the Golgi apparatus membrane to await assembled nucleocapsid for the budding process (Bardeletti et al., 1979, Hobman, 1989). Capsid protein of RV differs from the C of aiphaviruses in that it does not contain auto-9-  proteolytic capacity and it is therefore believed that the E2 signal peptide remains with the RV C protein (Suomalainen et al., 1990; McDonald et al., 1991). Although this explains the microscopic observations that capsid and assembling capsomere units are found in association with cellular membrane systems (Horzinek, 1981), the C protein and the assembled nucleocapsid have complete hydrophobic properties under physiological pH conditions (Mauracher et al., 1991). In SV, the C protein specifically interacts with viral RNA sequences (Weiss et al., 1989) and the encapsidation process is thought to be catalyzed by such an RNA-protein binding event in RV also. The assembled nucleocapsids can be readily detected in the cytoplasm and seen to bud into either the Golgi or from the plasma membrane (Oshiro et al., 1969) thereby completing the virus life cycle. Capsid protein has recently been shown to also bind to ribosome subunits (Singh et al., 1991) and such binding activity could possibly be involved in the control of the transition between protein synthesis and virion assembly.  1.2. Rubella Pathoaenesis and Patholov 1.2.1.History The first description of a disease induced by RV infection was described as “R6theln” by a German physician in the 18th century (in: Forbes, 1969), likely leading to the common name of “German Measles”. In 1866 the English-speaking medical establishment (maybe because of the annoyance of having to pronounce an umlaut), fmally followed the lead of the Scottish physician H. Veale in renaming this exanthematous rash and febrile disease to “Rubella” (Veale, 1866). To complicate matters, this disease obtained a third English identifier, following the attempt to classify acute exanthematous diseases into “First” through “Sixth” disease; Rubella, henceforth was known also as “Third disease” (Marcy and Kibrick, 1972). The viral etiology of rubella was confirmed in 1938 by human to human transmission studies (Hiro and Tasaka, 1938), thereby fulfilling Koch’s postulates.  1.2.2.Clinical and Immunological Features 1.2.2. LRubella a) Clinical Features The clinical features of this disease can range from a completely inapparent infection to the characteristic childhood presentation of low grade fever, adenopathy, malaise and exanthem. The exanthematous rash evolves -10-  rapidly and in most cases spreads from the head to the trunk, where by the second day it may appear as a pin point or reticular pattern. The rash usually disappears by day four causing no scarring. Other epithelial surfaces such as the conjunctivae or the oro-pharyngeal region may also show signs of inflammation. In more severe cases, the infection results in encephalitis and in rare cases (< 0.05%) severe encephalitis can have a fatal outcome (Center for Disease Control, 1975). In cases of post-pubertal infection, the disease generally has a more severe outcome. Especially in young women the disease is associated with articular complications ranging from acute arthralgia to chronic arthritis (Smith et al., 1987). This outcome will be discussed in more detail in section 1.3.4.. The common symptoms of Rubella can be readily confused with similar illnesses presenting with a maculopapular rash, malaise and adenopathy. In one particular study of patients having a rubelliform rash, adenopathy, generalized malaise and joint pain only 36% were confirmed by laboratory tests to have had rubella. The remaining patients were infected with parvovirus (7%) or had an unlcnown etiology (57%) (Shirley et al., 1987). Defmite diagnosis can therefore only be made following laboratory studies involving either virus isolation or serological analysis. Rubella is highly communicable with spread, most probably, occurring by droplet infection or direct mucosal contact. Mucosal epithelial cells and submucosal lymphoid tissue serve as the initial infection site from where the virus spreads to regional lymph nodes. Rapid viral replication in affected lymph nodes leads to enlargement and tenderness which may start 5-10 days prior to the outbreak of a rash (Green et al., 1965). An incubation period of 7-9 days precedes the onset of viremia and shedding of the virus from nasopharyngeal secretions and stool (Heggie and Robbins, 1969). The exanthem commences at 16-2 1 days following initial exposure and coincides with the appearance of IgG in the blood stream. As the rash onset coincides with the generation of anti-RV IgG it has been hypothesised that the exanthem is caused by immune complex mediated vessel injury (Heggie and Robbins, 1969). The viremic phase quickly subsides with the mounting of the specific immune response and in general virus can not be isolated from biological fluids one week after initiation of rash (Davies et a!., 1971). In cases of experimental RV infection or following vaccination these symptoms may occur up to a week earlier than natural infections (Anderson, 1949).  b) Immunological Features The initial antibody response to RV infection or RV vaccination is the 1gM response, which in most cases is —11—  1024 ELISA , 4%..  R A  100  64  HAI.  \ -0  -.-—..  76543211234  6  8  10  12  1  14 I  Days Before  Days After Onset of Rash  2  rm, 1 2 3  Months  10 I  LJ  Years  Figure 5. Clinical Features of Rubella. Patterns and kinetics of viral excretion, rash (shaded area) and antibody responses during primary Rubella. (HAl) Hemagglutination lithibiting antibody, (CF) Complement Fixing antibody, (rIgM) 1gM treated with reducing agent. (From: Cooper and Buimovici-Klein, 1985)  specific to the El protein (Partanen et al., 1985). This response is transient, lasting on average for one month (see Fig. 5) but cases have been reported where 1gM responses are detected for periods of months or years following initial Rubella infection (Al-Nakib et al., 1975). 1gM persistence is not necessarily correlated with specific pathology but the occurrence of persisting 1gM can make it difficult to confidently time initial RV infection by serological means. All remaining immunoglobulin classes are stimulated within weeks of infection including IgG, IgD and IgE (Salonen et al., 1985). The IgA response is felt to be transient, restricted to IgAl and restricted to C protein (Partanen et al., 1985; Stokes et al., 1986). However more recent studies have shown that the serum IgA response may be long lasting and directed to all three proteins of RV (Chapter 3; Zhang et al., 1992). The IgG response is by far the dominant serological response and is directed to all three structural proteins of RV. When analyzed by immunoprecipitation techniques it is observed that the majority of RV specific IgG is directed to El with significantly lower levels being directed to E2 and C (Katow and Sugiura, 1985, de Mazancourt et al., 1986). This holds true also if sera are analyzed by immunoblot or protein specific ELISA (Zhang et al., 1992,Chaye et al., 1992). The RV specific IgG response is predominantly made up by the IgG 1 subclass with small amounts of IgG3 and IgG4 (Sarnesto et al., 1985). The levels of IgG and its subclasses correlate well with HAl titres (Stokes et al., 1986) and with neutralizing -12-  antibody titres (Stewart et a!. ,1967) and it has been assumed that these responses play a major role in viral clearance.  The affinity of RV-specific IgG increases rapidly in the months following infection or vaccination  (Hedman and Rousseau, 1989) and these biologically active antibodies are maintained indefmitely if infection occurs in childhood and contributes to the long term resistance to RV reinfection (Wolinsky, 1990). Circulating immune complexes (CIC) containing RV-specific antibodies and antigen are frequently detected following infection (Ziola et al., 1983). The cellular response to RV is vigorous and in healthy responders is maintained at low levels over an indefinite period (Buimovici-Klein et al., 1977). RV-specific cellular responses are demonstrable by lympho proliferative responses, lymphocyte-mediated cytotoxicity and are MI{C restricted (Ilonen et al., 1986). Recent work has shown that El shows the highest reactivity in RV protein specific lymphoproliferative responses in adult humans, followed by C protein with weakest reactivities measured to E2 (Chaye et al., 1992). The predominance of both anti-El B- and T-cell responses in RV specific immunity have led to the conclusion that immunity to El is central to protection from viral reinfection. This is further supported by the almost exclusive distribution of neutralizing domains on the El protein.  1.2.2. 2.Congenital Rubella Syndrome a) Clinical Features The characteristically uneventful outcome of childhood rubella stands in marked contrast to the devastating outcome of in utero RV infection. The connection between matemal rubella in the first trimester of pregnancy and the high incidence of congenital malformations, especially cataract formation, was first recognized in 1941 by an Australian ophthalmologist in the months following a Rubella epidemic (Gregg, 1941). Following this initial description, the teratogenic effects of this virus were quickly established and instigated the global effort for vaccine development and production. Maternal RV viremia prior to conception is not associated with an infection risk to the fetus (Enders et al., 1988), however once gestation commences the risk of in utero infection becomes high, with few fetuses escaping infection if maternal Rubella occurs in the first month of gestation (South and Sever, 1988). The rate of fetal infection diminishes to an estimated 25% -30% following the first trimester (Miller et al., 1982). The severity  -13-  of birth defects and late onset sequelae strongly correlate with the gestational age of the fetus at infection. In general it is believed that the closer infection occurs to conception the more severe the outcome will be. Congenital malformations are expected if infection occurs prior to 16 weeks. Birth defects are rarely described if infection occurs later than 17 weeks; late onset sequelae such as neurological deficits or endocrine disorders (hypothyroidism or diabetes mellitus) are however commonly described in congenitally infected patients and may be attributed to RV exposure (Munro et al. 1987). RV readily infects placental tissue and virus can usually be isolated from this site (Alford et al., 1964). During early gestation, the placenta is observed to contain scattered foci of necrotic trophoblast tissue as well as vascular endothelium (Tondury and Smith, 1966). The invasion of the virus into the feto-matemal circulation allows for an embolic mode of transmission. This can explain the disseminated infection of the fetus affecting almost every organ system (Bellanti et al., 1965). The mechanism by which the virus interferes with organogenesis is not clear but several theories have been proposed to explain the observations of organ dysfunction and lowered cell numbers in organs which otherwise appear anatomically normal. It has been suggested that virus can directly damage cells or can persistently infect fetal cells and thereby interfere with the ontogeny of organs (Rawls and Melnick, 1966). A more indirect mechanism has been proposed to be mediated by soluble factors, secreted by RV infected cells, inhibiting regular mitotic division (Plotkin et al., 1967). More recently it has been described that cells, persistently infected with RV, do not react well to growth factors and produce little collagen in comparison to uninfected cells (Yoneda et al., 1986). The general observations of damaged larger vessels both in the placenta and the CR5 fetus can explain lowered cranial size and birth weight as well as ischemic necrosis observed in neural tissues of some CRS patients. Immune complex mediated injury may also contribute much to the morbidity and mortality observed in perinatal and late onset sequelae. Inflammatory damage, ranging from acute neutrophil infiltration to chronic inflammatory lesions, is commonly described in lung or the brain at autopsy in CRS patients (Desmond et al., 1967). Increased loads of CIC, containing anti-RV IgG, have also been described in CRS patients with active clinical complications and have been interpreted as being indicative for the systemic, low grade replication of RV in these patients (Coyle et al., 1982). Table 2 lists some of the common clinical features of CRS. Viral persistence is one of the hallmarks of the CRS. Virus may be isolated from most organs at birth and RV is actively secreted in urine, stool and naso-pharyngeal secretions in more than 80% of CRS patients in their -14-  Feature  Incidencea  Intrauterine growth retardation (low birth weight) Thrombocytopenic purpura Hepatosplenomegaly Meningoencephalitis Bone lesions (radiographic) Large anterior tontanelle Adenopathy, generalized Hepatitis Cloudy cornea Hemolytic anemia Pneumoraa due to rubella Myocarditis due to rubella Deafness. sensorineural Central language disorders Mental retardation Behavioral disorders Spastic diplegia Patent ductus arteriosus Pulmonic Stenosis Cataract (and microphthalmia) Retinopathy Glaucoma Severe myopia Inguinal hernia Cryptorchidism  TC TC TC TC TC TC TU TU TU TU TtJ TU PC (DU) PDC PDC PDC PC PC PC (?DU) PC PC PU (DU) PDU PU PU  a Transient (T): permarwnt (P); developmental (D); common (C); or uncommon (U).  Table 2. Clinical Features of Congenital Rubella (From: Cooper and Buimovici-Klein, 1985).  first month of life (See Fig. 6). Thereafter, the rate of viral excretion diminishes and is only rarely detected once the patient has reached the age of 2 years. However it is felt that RV persists in organs of most CRS patients as virus has been isolated from patients 20 years and older (Menser et al., 1967; Weil et al., 1975). With the advent of PCR technology a more accurate estimate of the incidence and preferred organ distribution of persisting virus in this patient group may be forthcoming. Persistence of RV in these infants has remained a paradox, as this virus readily resists clearance despite the presence of maternally acquired immunity as well as the autologous anti-RV 1gM response. As CRS patients do not exhibit generalized immunosuppression after infancy and are capable of responding to the routine pediatric vaccines it has been proposed that a form of RV-specific tolerance exists. In several animal models it has been shown that in utero infection lead to immunological tolerance resulting in the persistence of the virus.  -15-  I00 90  71/85’84%  80 70 5OI8I62% 60 50  ‘z 40 26/8033% 30  w C-,  20 II/s8lrh 10 C  0/20’O% 0-I  9-12 5-8 MONTHS AGE Of INrANTS AND CHILDREN -4  13-20  / YEARS  Figure 6: Incidence of virus shedding by age in CRS patients (Cooper and Krugman, 1967).  Examples such as LCMV in mice (Traub, 1938), Borna agent in rats (Hirano et al., 1983) and Border disease virus (BDV) in sheep (Barlow, 1983), have described that infection during the early ontogeny of the immunesystem can lead to inmunologic non-responsiveness of the fetus leading to a lifelong, low grade persistent infection of the individual. Immunological tolerance to these agents is rarely complete but rather a state of split tolerance is established.  In the case of LCMV or the Borna agent, cytotoxic T cells (Tc) and delayed type  hypersensitivity (DTH) reactions to viral antigens are not detectable. However helper T cell (Th) cell responses are believed to remain intact as IgG specific to the three LCMV structural proteins is produced; this hypothesis is supported by the observation that adult athymic nude mice, infected with LCMV, only respond with a transient 1gM response (Buchmeier et al., 1980). However the IgG antibody in the tolerant animals is of low affinity and shows weak or no in vitro neutralization activity. The similarities between the immunity of LCMV tolerized mice and the CRS patients will be further discussed in section 1 .2.2.2.c.Other explanations for viral persistence may include cryptic viral infections in immunologically preferred sites such as lens, salivary glands, cartilage or neural tissue. Virus might re-emerge from these sites following the waning of passively acquired IgG immunity. A more plausible explanation is that the presence of in vitro neutralizing antibody, transferred from mother to fetus, may be of no consequence for the replication of virus in vivo. This suggests that the cytotoxic responses  -16-  are most important in protective immunity and as the cellular response is not transferred from mother to fetus, the viral infection can not be cleared until an autologous cellular responses is mounted. Recent evidence has shown (Chaye et al., 1992) that viral-specific T cell responses to the El protein are significantly reduced in CRS patients and that this perturbation in the viral specific T cell repertoire may lead to deficiency in viral clearance.  b) Immunological Features The congenitally infected fetus actively produces RV-specific 1gM at 15 to 20 weeks of gestation (Daffos et al., 1984). A positive RV 1gM titre in cord blood is therefore interpreted as serologic evidence for intrauterine infection with RV. However this does not correlate with the development of birth defects nor does the absence of virus-specific 1gM preclude this syndrome (Enders and Jonatha, 1985). The newborn carries passively transmitted maternal IgG, which in the first year will be removed at an average  half-life rate of 21 days  (Hamilton, 1987). IgG of all subclasses can cross the placenta (McNabb et al., 1986) and CRS patients or their cord blood normally test positive for HAl and neutralizing IgG, reflecting the passively acquired maternal IgG (Sato et al., 1979). Nevertheless neither the maternal IgG nor colostrum/breast milk derived IgA is capable of eliminating viral replication. Over the first year, the infants mount their own RV-specific antibody response, switching from the endogenous 1gM response to IgG and IgA. Because of the reported long term persistence of RV, it has long been speculated that a form of immunological tolerance must exist to allow viral persistence. This RV-specific tolerance can not be complete as CRS patients produce a RV-specific IgG to both envelope and capsid proteins and by inference must therefore also possess Th cell responses. Much controversy exists in the literature on the quantitative levels and protein-specific responses of both the humoral and cellular RV responses in the CRS patients. Initially it was observed that CRS patients quickly lost HAl antibodies in their first year of life indicating a specific loss of anti-El antibodies (Cooper et al., 1971; Ueda et al., 1987). It was ,  then proposed that tolerance in RV patients was indicated by their inability to produce IgG to RV El, while maintaining reactivity to E2 and C. A 1985 study comparing the protein-specific antibody reactivity of CRS- and control patients using immunoprecipitation technique showed that CRS patients had significantly lower levels of El- and elevated levels of E2-specific IgG (Katow and Sugiura, 1985), although these fmdings were not confirmed by others (de Mezancourt et al., 1986). The affinity of antibody directed to RV has been found to be low in CRS children, in comparison to age-matched RV sero-positive controls (Fitzgerald et al., 1988). Both -17-  lowered affinity and non-responsiveness to key proteins or epitopes may explain viral persistence by deficient antibody mediated viral clearance mechanisms. The study of RV-specific T-cell responses, measured by 3 H-thymidine uptake or interferon release, has been described to be selectively reduced in CRS patients (Buimovici-Klein and Cooper, 1985). Moreover, these authors report that the degree of tolerization is dependent on the gestational age at infection, with infection in the first 16 weeks leading to the most pronounced reduction in T cell responses to whole RV. Others have reported that CRS patients have reduced ratios of CD4 + /CD8 + cells, as well as an overall reduction of activated Th cells as indicated by significantly lower Ia positive cells than age-matched control patients (Rabinowe et al., 1986). Although suppression of RV-specific responses are suggested in CRS patients, a rudimentary Th cell response must exist to support anti-RV IgG production to the viral antigens.  Recent  evidence from studies using recombinant RV proteins or RV peptides in lymphoproliferative assays, suggests that only responses to the El protein are selectively suppressed, whereas responses to the E2 and C protein are normal or elevated in CRS patient groups (Chaye et al., 1992; Ou et al. 1992). This thesis will propose a model which explains the selective loss of anti-El T cells in the presence of IgG specific to all three RV proteins for the CRS patient group.  1.2.3 .Tolerance following Intrauterine Virus Infection Both RNA and DNA viruses have developed a variety of means of producing persistent infections in their hosts (Table 3). Indeed it can be speculated that the earliest viral pathogens of the emerging human species had to be capable of inducing persistent infections as the human hosts were distributed in small pockets of population which rarely intermingled. This is supported by epideniiological studies on isolated populations in the Amazon, where it has been shown that the dominant viral infections are caused by persistently infecting viral species (Black et al., 1970). Only with the onset of organized agriculture did human populations reach basic community sizes which were large enough to support viruses which do not establish persistence in the host (ie: influenza, smallpox) (Fenner et a!, 1974). Immunological tolerance to a virus (defined as the virus-specific refractoriness in the immune response following prior exposure to this virus) may be one mechanism which viruses have evolved to establish persistence in the vertebrate host. Animal models such as the neonatally LCMV infected  -18-  Mechanisms of Viral Persistence  Avoiding Immunological Surveillance 1)  Lowering or avoiding the presentation of antigen on MHC I and II molecules a) Down-regulation by direct virus infection [Adenovirus (Tanaka et al., 1985)]. b) Infection of cells devoid of IvIHC-I, eg: neural cells [HSV I (Fitzpatrick et al., 1991)].  2)  Infection of cellular constituents of the immune system [HIV (Hirsch and Curran, 1990)].  3)  Generalized Immunosuppression [Measles virus (McChesney and Oldstone, 1987), FUV (Hirsch and Curran, 1990)].  4)  Induction of Tolerance [Lymphocytic Choriomeningitis virus (LCMV) in mice (Buchmeier et al., 1980); Rubella virus (Coyle et al., 1982; Mauracher et al., 1992c)].  5)  Infection of an immunologically privileged site [Papilloma viruses (Carson et al., 1986)1.  Alter Replication and Transcription of Virus 1)  Production of Defective Interfering Particles [Adenovirus, Rhabdovirus (Huang, 1988)].  2)  Generation of immunological variants [liv (Ruscher et al., 1988)].  3)  Loss of viral protein from the surface of the infected cell. a) Capping and antigen modulation by anti-viral IgG [Measles (Fujinanii and Oldstone, 1984)]. b) Budding of virion into Golgi or ER [HIV (Narayan et al., 1988); Rubella (Hobman, 1989)].  4)  Insertion of genomic information, with transient expression (latency) [Herpes viridae (Fitzpatrick and Bielefeldt, 1991)].  Other Mechanisms 1)  Poor antigenicity [Scrapie and other agents causing spongiform encephalopathies (Chesebro, 1985)].  2)  Continuous infection of host or population by vector mediated infection from exogenous pool or zoonotic infections [Dengue virus (Burke et al., 1987)].  Table 3. Mechanisms of Viral Persistence. For each mechanism, an example of a human or other vertebrate viris is given. Potential mechanisms for RV persistence are listed, however this virus may use other mechanisms in its persistence.  mouse, which leads to a life-long infection (Traub, 1938), have led to the concept of immunological tolerance (Bumet and Fenner, 1949) and the formulation of Burnet’s clonal selection theory. These have also inspired extensive studies of tolerance induction by LCMV infection in the mouse model as well as in other virus/animal -19-  models (Nash, 1985). The persistently LCMV-infected mouse was initially reported to be completely tolerant to the virus, but it was later reported that these animals were capable of producing antibody to all three LCMV proteins. These however were found as deposited IC in spleen and kidney (Oldstone and Dixon, 1969) where they could induce IC-mediated immunopathology. These virus-specific IgGs were found to be non-neutralizing ai-id of low functional affinity (Buchmeier et a!., 1980). The Tc and DTH responses are absent in these animals. Although studies on in utero infected animal models has clarified many concepts in immunological tolerance it was the introduction of the transgenic mouse model which allowed a more detailed interpretation of tolerance induction in both B and T cells. With the controlled expression of transgenic proteins in specific organs the role of both periferal and central tolerance could be investigated, as well as the time point in lymphocyte ontogeny when tolerization occurred.  Once it was recognized that the Th cell was an essential regulator of B cell  progression, the issue was raised whether B cell tolerance was essential for self-education or if tolerance of the T cell responses was sufficient in the regulation of antibody responses.  Mice which have been made double  transgenic for hen egg lysozyme (HEL) and for high affinity anti-HEL antibody have provided evidence for B cell anergy (Goodnow et al., 1990), independent of T cell control. The model using MHC transgenic mice has shown that exposure of the pre-B cell to high levels of high density toleragen can lead to clonal B-cell deletion in the bone marrow (Nemazee and BiYrki, 1989). On the other hand studies using a peripherally expressed, soluble transgene product (human insulin) have shown that B cell anergy does not occur but that antibody nonresponsiveness is solely due to peripheral T cell tolerance (Whiteley and Kapp, 1989). Recent work using a transgenic G membrane protein of Vesicular Stomatitis virus (VSV) has also provided supportive evidence that B cell tolerance to a monovalent, periferally expressed protein is regulated by Th cells (Zinkernagel et al., 1990). In these transgenic mice B cell tolerance remained complete when animals were challenged with recombinant G protein but tolerance was broken if whole VSV was used as an immunogen. Presumably the anergic anti-G protein B cells could be stimulated into IgG production by T cells specific to other viral proteins. Toleragen concentration, tissue distribution and multivalency clearly play a key role in B cell tolerance, with exposure of the immature B cell to multivalent antigens (ie: MHC) leading to clonal deletion.  In encountering soluble  monomeric proteins or membrane proteins of low copy number clonal anergy may occur but it now seems evident that the fate of the B cell is controlled by the underlying Th cells repertoire (Nossal, 1991). Suppressor  -20-  T cells in B cell tolerance (Theopold and Köhler, 1990) and the anti-idiotype network theory (Jeme, 1974) may also play a role in the regulation of self-recognition but are beyond the scope of this introduction. Many similarities exist between the tolerant state of mice following neonatal or intrauterine LCMV exposure and humans infected with RV in utero during the first trimester of gestation. In both cases there is an IgG response which is of low affinity and which lack in vitro neutralizing activity. IC formation and deposition followed by IC-mediated inflammation has been described in both models.  Whether CRS patients show a  complete lack of virus-specific Tc and DTH reactivity, as is the case in the murine model, remains speculative. Recent data suggests that lymphoproliferative responses to the E2 and C protein of RV remain intact in CRS patients and it is likely that Te responses to the these minor T cell antigens are present (Chaye et a!., 1992). It is proposed that tolerance is induced in CRS patients to the immunodominant El protein, rather than to all viral proteins as has been reported in the LCMV model.  1.3 Rubella Vaccine 1.3.1.History Following the descriptions of the devastating outcome of congenital rubella (Gregg, 1941), a worldwide effort was undertalcen to develop measures to prevent fetal RV infection. Initially, gamma globulin was administered to pregnant women recently exposed to rubella (described in Chase, 1984) but it was not until 1962 when the etiological agent of rubella was finally isolated by two independent groups (Weller and Neva, 1962; Parkman et al., 1962) that vaccine development could be initiated. Between 1965 and 1967, several live attenuated RV strains were developed and tested as vaccines. Of these HPV 77/DE 5 and RA 27/3 have been licensed in Europe and North America.  Currently the RA 27/3 strain of RV which is passaged on human diploid  fibroblasts (Plotkin et al., 1969) is the commonly used vaccine strain in Canada and the USA. This vaccine was not licensed for use in North America until 1979 but has been used in many European countries from 1969 onward.  With the development of less virulent vaccine strains the public health offices of North America  changed their recommended vaccine scheduling.  The USA health policy in 1969 recommended universal  vaccination of children at age one year and a one time vaccination effort of children up to age 12 (Ilimnan et al., 1983). This vaccine policy was in effect until the late 1970s when it became clear that the incidence of CRS  -21-  Vaccwe Licensed  30  —  r— . 5 N  28  ——  Rubella, lolal Rubella. aIS year olds  14  Congenilal rubella+  \  26  \ \  24 22  12  20 0 a g18  10  12  6  c)  (I,  a 10  a s_I  0  8 a-  4  /‘“  /  6  \  S.  •_.—‘  —..  5/  .5’,.  4  2 -—  -.  S.  2 -  0 1966  67  ‘68  •69  70  ‘7 172’7374 Year  0 75767778  79  ‘80  Figure 7: Incidence of reported rubella and CRS cases in the USA, 1966-8 1 (Hinman et a!., 1983).  and reported rubella cases in females of child bearing age had not dropped to levels that were initially expected (Figure 7). In 1978-1979 it then became common practice to immunize females of child bearing age and to advise post partum immunization of RV seronegative mothers (Preblud, 1986). This change in vaccine policy led to a significant drop in reported cases of CRS and has remained vaccine policy in both the USA and Canada.  1.3.2.Biology and Immunology The mechanism of RV attenuation has remained unclear although both wild type strains and attenuated vaccine strains have been sequenced (Zheng et al., 1989). Attenuated strains vary minimally in protein sequence, with only 5 amino acid substitutions occurring in the structural polyprotein of the HPV 77 RV vaccine strain compared to wild type isolates. The low virulence and slow replication rates of vaccine strains are thought to be due to a deficit in cell attachment and entry (Nakhasi et al., 1989), although other mechanisms of attenuation (eg: host protease dependence) can not be ruled out. Immunity to vaccine strain virus is believed to be life long (Plotkin et al., 1973). After MMR vaccination at 12 months or older, 95% of children seroconvert by HAl technique and this immunity has shown to be protective for at least 15 years in 90% of vaccinees (Robinson et al., 1982).  -22-  1.3.3.Side Effects of RV Vaccination Controversy over the safety of rubella vaccine has only recently become of public concern. In the early seventies, neuropathies (Kilroy et al., 1970) and transient arthritis (Spruance and Smith, 1971) were describe in children receiving the HPV 77/DE 5 strain and led to the withdrawal of this strain and the licensing of the RA 27/3 strain in the USA and Canada (in Howson et al., 1992). Although the RA 27/3 strain has been shown to cause minimal side effects in children (Polk et al., 1982) it is still associated with significant adverse effects in adolescent and adult female vaccinees. Acute arthritis is estimated to occur in 13% -15% of adult RA 27/3 vaccine recipients (Reviewed in: Howson and Fineberg, 1992). Case reports of severe and chronic arthritis have also been reported following the administration of RA 27/3 in adult females (Tingle et al., 1986). In the pre licensure trials of the RA 27/3 vaccine this strain was determined to be safe for use in children. Subsequently these results were interpreted to hold true for adults as well, leading to the extension of vaccine administration to women of child bearing age without further evaluation of potential side effects (Tingle, 1990). Ongoing double blinded, randomized and placebo-controlled clinical trials of rubella vaccine in adult seronegative women at the BC’s Children’s Hospital Vaccine Evaluation Centre will provide a more definitive estimate of the incidence of acute and chronic forms of arthritis following RV vaccination. The pathogenesis of Rubella Associated Arthritis (RAA) has remained unresolved (Ford et al., 1986). Rubella can be isolated from synovial fluid and in peripheral lymphocytes in individuals with RAA, in the years following RV infection or immunization (Ogra et al., 1975; Chantler et al., 1985). Such fmdings have led to the proposal that RV can persist in the cells of the synovial fluid or the synovial membrane.  RAA is most  commonly associated with peripheral joints (Sauters and Utsinger, 1978) which are at temperatures of 5 #- 6 #C below the body’s core temperature.  This is of significance in light of reports that RV readily establishes  persistent infections of synovial cell cultures at 32#C but not at 36 #C (Cunningham and Fraser, 1985). Recently others have shown that wild type and vaccine strains of RV replicate to different extents in human synovial cell lines or in human synovial membrane organ cultures (Mild and Chantler, 1992). They showed that strains which have been reported to have a higher incidence of RAA are also characterised by a high degree of synoviotropism. Although it is likely that RV may persist in local tissues it remains unclear if this non-lytic virus is sufficiently virulent to cause direct damage in the joint or whether persistent RV antigen is connected with the initiation of immunopathology. It thus remains an enigma why some individuals mount a response which -23-  leads to rapid viral clearance and uncomplicated outcomes in Rubella while others, especially adolescent and adult females, mount immune responses which may not clear RV in localized organs and lead to the development of acute or even chronic arthritis. The pathogenesis of infectious arthritis has been summarized as “..a chronic T-lymphocyte and macrophage -  dependent response to foreign or autoantigens present in the synovial tissue. “(Decker et a!., 1984), however the question remains what initially attracts these cells to the synovium. A role of CIC and their deposition in the synovial space has maintained itself as an attractive hypothesis for involvement in the pathogenesis of RAA as well as in the manifestations of other forms of infectious arthritis (Inman et a!., 1987). IC are involved in inflammatory responses of CRS patients (Tardieu et al., 1980) and are most likely responsible for the vessel damage which presents as the rubelliform rash in acute Rubella (Heggie and Robbins, 1969). It is feasible that RV antibody and low grade locally persisting viral antigen may result in IC deposition and lead to the initiation of an inflammatory response. Studies regarding the presence of immune complexes have shown that levels of cic in RAA patients are not significantly increased in comparison to healthy Rubella patients (Singh et a!., 1986), however others have shown that cic containing anti-RV IgG are increased following RV vaccination (Coyle et al., 1982) but this study did not compare anti-RV containing CIC in healthy and RAA patients. A study investigating the levels of CIC containing anti-RV IgG or RV antigen as well as IC in the joint space may claril’ the role of IC-mediate injury in RAA. Reactivity of the virus-specific immunity with self-antigens has been proposed (Lund and Chantler, 1991), describing cross-reactivity of rubella specific epitopes with an unidentified protein derived from synovial epithelium. The hypothesis of molecular mimicry remains of interest and studies showing cross-reactivity of viral specific T  -  and B-cells may shed further light on the pathogenesis of RAA.  A recent model for the pathogenesis of RA has involved the role of undefmed infectious agents as superantigens (Paliard et a!., 1991), and the role of RV in this context is being currently investigated. Interestingly the RV capsid molecule shares structural similarities with the well defmed S. aureus enterotoxin superantigen (Marrack and Kappler, 1990), in that both contain an 8 membered b-pleated barrel as their core structure. Evaluation of the proposed immunopathological mechanisms in RAA requires a good understanding of the  -24-  “normal” adult immune response in order to be able to define differences in the immune responses of RAA patients which may be involved in the induction of the inflammatory processes. Furthermore similarities may exist in the immune repertoire of CRS- and RAA patients as both patients group may develop inefficient immune responses which lead to viral persistence.  1.3.4.Viral Vaccine Efficacy The recent controversy in the public media and medical literature (reviewed in: Howson et al., 1992) over RV vaccine and its induction of infrequent debilitating side-effects, has obscured the benefit that this vaccine program has brought to the public. The incidence of CRS in the USA has dropped from 106 cases in 1979 to 20 cases in 1985 with a continuing downward trend, so that CRS is hoped to be on the verge of elimination in North-America (Cochi et al., 1989). In the case of seronegative females, it should be remembered that the incidence ofjoint-manifestations following RA 27/3 administration is significantly lower than following wild RV infection, to which these vaccinees would have been susceptible (Tingle, personal communication).  1.4. Laboratory Diaanosis of Rubella Virus Infection In order to initiate the comparative study on the humoral immune responses to RV and RV proteins in congenitally- and post natally infected individuals, new assays had to be developed and established assays had to be adapted for application to the individual RV antigens. This section will review the most commonly used assays for the detection of RV or RV-specific immunity.  1.4. 1.Virus Isolation The defmitive diagnosis of RV infection is achieved by the isolation of this virus by tissue culture techniques. The virus will infect a broad range of cell lines. Virus samples can be isolated from pharyngeal swabs or nasopharyngeal secretions at the time of onset of the exanthem.  In the congenitally infected infant, virus is  readily isolated from stool, urine, nasopharynx, blood and cerebrospinal fluid. Virus can also be isolated from blood, synovial and breast milk lymphocytes (Chantler and Tingle, 1982), but these preparations do not yield virus in all cases and are not used for routine clinical diagnosis. Excretions or cell preparations are most commonly taken up in solutions of physiological salt solutions containing 1% BSA or 5-10% fetal calf serum and -25-  transferred immediately onto the cell line of choice for viral propagation. Two commonly used cell lines, African Green Monkey kidney cells (Vero) and Baby Hamster kidney cells (BHK-12) are used for growing high yield virus. The virus causes little cytopathic effect in these cells and virions can be readily isolated from tissue culture supernatants following the second day of infection. Other cell lines such as primary African Green Monkey kidney cells have been used to detect RV by interference with echovirus 11 replication and Rabbit kidney cells (RK-13) are extensively used for RV titre determination as this cell line exhibits a detectable cytopathic effect (Nawa, 1979). With the advent of sensitive molecular probing assays, direct detection of RV genome may replace viral isolation by tissue culture.  Reports of RNA hybridization for the detection of RV RNA in  chorionic villus samples (ITo-Terry et al., 1986) and the use of PCR technology in direct RV detection (Eggerding et al., 1991) indicates the direction of future virus detection. Serological testing for antigen specific antibody responses, remains the primary diagnostic means for RV infection. Serological techniques are independent of the narrow time interval when virus can be isolated and thereforee can be used retrospectively to confirm and time RV infection.  The following paragraphs will  introduce the serological techniques which have found application in RV serology and some of the problems associated with their application and interpretation.  1.4.2.Neutralization Assay The neutralization test (NT) is one of the oldest assays in viral diagnosis and serology, dating back to the early work on vaccinia virus (Stemberg, 1898 reviewed in: Lennette et al., 1989). Stemberg observed that certain sera were capable of rendering vaccinia virus non-infective to a susceptible host, and termed this activity “neutralization”. This serologic test has been the “goldstandard” against which other serological techniques have been measured.  In rubella serology this assay has been in use since the isolation of this agent and directly  measures the individuals capacity to neutralize RV infection in vitro. However because NTs are labour intensive they have not found much application in routine diagnosis and are mainly employed in research laboratories for the standardization of new serodiagnostic tools (Steward et al., 1967) or in the characterization of RV specific monoclonal antibodies (Waxham and Wolinsky, 1985b; Brush, 1989). In the case of RV neutralization it was soon realized that serum components other than immunoglobulins had the capacity of neutralizing the virus.  -26-  These have been determined to be complement, b-Iipoproteins (Clarke and Cassals, 1958) and phospholipids (Mastromarino et al., 1990). Such non-specific inhibitors have to be removed by pretreatment of sera with 57°C incubation followed by kaolin treatment or by precipitation with manganese chioride/heparin (Kawano and Minamishima, 1987) prior to the determination of neutralization titres.  1.4.3.Hemagglutination Inhibition (HAl) Assay Many viruses from varied families are capable of agglutinating erythrocytes from a variety of species. This hemagglutination (HA) phenomenon forms the basis for the serodiagnostic test of antibodies which bind to the virus and thereby inhibit HA (HAl). RV was observed to bind to day-old chick erythrocytes as well as white goose erythrocytes and this interaction can be selectively inhibited by antibodies as well as the non-specific inhibitors described above.  The HAl assay had been extensively tested and compared for sensitivity and  specificity to the NT (Lennette et a!., 1967; Field et al., 1967) and has been found to show a good correlation with NT in human test populations. The standard protocol (U.S. Dept. of Health, Education and Welfare, 1970) recommended pretreatment of sera with manganese chioride/heparin and adult chicken erythrocytes for the removal of non-immunoglobulin inhibitors of HA and has set a HAl titre of 1:8 as the cut-off for protective immunity. The HA! assay will provide results in one day and demands a fraction of the costs of the NT; therefore HAT was quickly adapted by most laboratories as the mainstay of rubella serology from the mid 1960s to 1980s, and provided the reference standard to which novel immunoassays had to be compared (Herrmann, 1985). The molecular mechanism of the HA interaction is not completely understood in that the receptor on the erythrocyte is not defined. The HA activity on the virion on the other hand has been studied extensively. HA activity resides on the El protein of RV and is dependent on intact conformation as well as complete glycosylation (Qui and Gillam, 1992). Binding sites for HAl antibodies have been mapped to a 40 amino acid region on El. These are in close proximity to the defined neutralizing epitopes but do not correlate in all cases (Terry et a!., 1988; Waxham and Wolinsky l985a).  1.4.4.Enzyme-Linkedlinmunoassays (EIA) Enzyme-linked immunoassays have found much application in serodiagnosis since the large scale availability of enzyme-conjugated monoclonal and polyclonal antibodies directed to immunoglobulins first described in 1969 -27-  Figure 8. Hemagglutination-Inhibition (HAl) Test. Erythrocyte sedimentation [ A] correlates to the presence of antibodies which have bound to El protein and inhibited the cross-linking of cells by the virion. Hemagglutination [ ©] means the absence or the diluting out of antibodies, as the virus exerts its agglutination properties. Each row represents a two-fold serum dilution from 1:8 to 1:512. Top row illustrates a negative serum, the remaining three are positive with titres of 1:64,32 and 16 respectively (From: Lennette et al., 1989).  (Avrameas, 1969; Engvall and Perlmann, 1971). Conjugation procedures and substrate systems have improved dramatically in the last 15 years and have led to the large scale acceptance of this technique.  Although  fluorescent- and radiolabelled immunoassays find application in specialized fields, the enzyme conjugates, mainly alkaline phosphatase and horseradish peroxidase, have emerged as the most versatile and applicable to immunodetection  and diagnosis (Lennette et al., 1989).  For RV serology in most clinical and research  laboratories, EIAs have replaced the older HA! and NT assays because of the increased sensitivity and economy of the EIA (Steece et al., 1984). The two following paragraphs will briefly review two applications of RV specific EIA which are central to this thesis: the enzyme-linked immunosorbent assay (ELISA) and Western blot. Enzyme-Linked linmunosorbent Assay (ELISA) The use of ELISA in RV serodiagnosis was initially described in 1977 (Bidwell et al., 1977). By the early 1980’s it had become the main laboratory diagnostic test for rubella.  As this test allows the determination of levels  of all immunoglobulin classes, it lends itself to automation and can be readily scaled up to screen large populations (Leinikki et al, 1978; Buimovici-Klein, 1980). A world-wide serum reference system has allowed for the uniform and standardized interpretation of ELISA results (WHO, 1971).  A good correlation was  observed between ELISA seropositivity and positive HAl titres (Shekarchi et al., l981;Herrmann. 1985: Stokes  -28-  et al., 1986) in normal patient populations while others have found this correlation to be poor in CRS patients (Hancock et al., 1986). The commercially available whole RV EUSAs measure antibodies directed mainly to surface RV epitopes of the envelope proteins and do not necessarily measure antibodies which are functionally significant for in vitro neutralization or which provide the host with protective immunity. This stands in contrast to functional antibody tests, such as HAl and NT, which measure antibody to a narrow range of epitopes on the El protein.  1.4.4. 2.Westem Blotting The electrophoretic transfer of proteins onto nitrocellulose from SDS-PAGE gels and the detection of protein by enzyme-conjugated antibody was first described in 1979 (Towbin et al., 1979). Although Western blotting most commonly employes enzyme-conjugated detection antibodies (immunoblot)  the use of flourescing  antibodies have come into use over the last years. In the case of RV serology, Western blotting allows the further dissection of the humoral immune response by determining the protein specificity of immunoglobulins. However disadvantages inherent to this technique (ie: protein denaturation, long processing time) have given it few applications in large scale serology, with the exception of AIDS serology (Steckelberg and Cockerill, 1988), so that this technique has remained a tool largely used in research laboratories.  Serological techniques for the analysis of Rubella immunity have evolved as a consequence of the demand for assays capable of rapidly, repetitively and cost-effectively handling large numbers of samples. Today, ELJSAs have filled this demand and provide a sensitive and specific test for the presence of anti-RV imniunoglobulins. These assays are therefore of use in the diagnosis of previous virus exposure  -  but do not provide a final answer  as to whether the detected immune response will protect an individual from reinfection.  This problem is  illustrated well by CRS patients who test strongly positive on RV EUSA but remain susceptible to reinfection (Cooper et al., 1971).  1.4.5.Affinity Assays The above discussed RV serology assays have been developed to measure RV- or RV protein specificity of  -29-  immunoglobulin classes, their quantitation and to some extent their biological function but provide limited information about the affmity of these iminunoglobulins. The high affmity of the immunoglobulin molecule to its antigen is an important functional parameter and is ultimately involved with antigen-specificity as well as the biological activity of the molecule (Table 4). As the terms affinity and avidity will be used extensively in the text and have often been used interchangeably in the literature, the brief defmitions below will help in the clarification of these terms. The term “Affmity”describes the quantitative interaction of hapten and antibody. The Law of Mass Action can be applied to this interaction and the stability of the antibody:hapten complex can be expressed and experimentally defined as the association equilibrium constant, K.  For an accurate  experimental determination of the K, both antigen and antibody should be present in pure and homogeneous form. The determination of K is therefore most precise in a system involving monoclonal antibodies interacting with monovalent or haptenic antigens (Steward and Steensgaard, 1983). The term avidity or “functional affinity” is used to qualitatively describe the strength of antigen: antibody interaction. Although directly dependent on the affinity of the antigen binding site with the antigen, it is strongly influenced by antigen valence and immunoglobulin class. This can be best explained by a set of hypothetical IgG and 1gM molecules with identical antigen binding sites for a multivalent antigen.  Both molecules have identical affinities however the 1gM  molecule will bind more avidly as it can combine with more sites on the antigen. It has been realized for many years that the affinity of antibodies progressively increases following immunization and this phenomenon has been termed affinity maturation.  The mechanism by which such an  increase of the average affinity occurs in response to immunologic challenge has received much attention and several underlying molecular mechanisms have found acceptance: somatic mutations (Milstein, 1987), antigenic selection (Allen et al., 1987) and a unifring hypothesis, combining the latter two and proposing hypermutation of the IgG V region on a DNA level during the G 0 phase prior to B-cell progression (Manser, 1990). The affinity maturation of anti-RV IgG responses following primary rubella infection occurs over a three to four month period (Hedman and Rousseau, 1989), with antibody affmity increasing by an estimated 100 fold. This range of increase in affmity is supported by studies in rabbits using DNP-lysine antigen, indicating that IgG from the acute response rose in affinity from 5x10 5 to 108 lM’ in the months following antigenic challenge (Eisen and Siskind, 1964). The well defined increase in the affinity of anti-RV IgG has been recently been exploited to determine the time point when RV infection occurred (Hedman et al., 1989). This is of significant -30-  Reaction  Reference  Passive hemagglutination  (Levine and Levytska, 1967)  Complement fixation  (Fauci et al., 1970)  Hemolysis  (Warner and Ovary, 1970)  Immune elimination of antigen  (Alpers et al., 1972)  Virus neutralization  (Blank et al., 1972)  Damage to DNP-sensitized liposomes  (Six et a!., 1973)  Protection against bacterial infections  (Ahlstedt et al., 1974)  Blocking antibodies  (Adkinson et al., 1979)  Hemolysis typing  (Hedman et al., 1989)  Table 4: Biological Reactions in which High Affmity Antibodies are Superior to Low Affinity Antibodies.  diagnostic importance in determining intrauterine exposure if recent RV infection during early pregnancy is suspected as the determination of affinity will provide a good estimate if Rubella occurred before or after conception. Antibody affmity considerations are also of importance in the study of immunological tolerance. The LCMV mouse model of tolerance was initially described to show complete immunological tolerance following neonatal virus inoculation (Traub, 1938), but was later shown to be a form of split tolerance. Cytotoxic responses were lacking in these animals (explaining persistent virus replication) but IgG responses to all three viral structural proteins were present. In most animals LCMV-specific antibodies were found as CIC or IC deposited in the reticuloendothelial system or in the kidneys. These antibodies were of low affinity in comparison to those found in adult responding mice and had no in vitro neutralizing activity. It remains unclear if the decrease in antibody affinity during immunological tolerance is due to preferential deletion of high affinity clones or if the lack of Th cell results in inefficient stimulation of B-cell clones in the germinal centres. However recent evidence indicates  -31-  that high affinity responders to self or toleragen are selectively purged (Nossal, 1991). Predominance of low affinity responses to a replicating antigen is therefore an indication that split immunological tolerance has occurred. A wide range of methods are available for the determination of antibody affinity. Equilibrium dialysis and animonium sulphate precipitation have found much application in studies using monoclonal antibodies but are of little use for the determination of functional affinities of sera to complex mixtures of antigens such as the Rubella virion. Two methods will be introduced here which will be extensively described in the body of the thesis. .Chaotropic Elution Assays Antigen-antibody interactions can be disrupted by exposure to pH extremes, chaotropic agents and detergents. Studies on immune complexes described that low affmity antibodies could be readily dissociated by chaotropic agents, whereas high affmity interactions were resistant to these conditions and could only be dissociated by conditions leading to the denaturation of both antigen and antibodies (Hogben et al., 1986). Protein denaturants could therefore be used to selectively disrupt the binding of low affmity antibodies by choosing the reagent concentration and the exposure time of the immune complex to the denaturant. Urea acts as a denaturant by interfering with hydrophobic interactions of polypeptide chains (Kamoun, 1988) and exposure of immune complexes for short periods to 8M urea, or other denaturants such as 4M isothiocyanate, leads to selective elution of low affinity antibodies from solid phase bound immune complexes. This can provide a measurement of the average affinity of serum antibodies by calculating the amount of total immunoglobulin which have resisted the chaotropic elution (elution ratio [ER(%)].  This ratio, also referred to as the affinity index, is  believed to be related to the average affmities of the circulating immunoglobulins directed to the antigen system, as these techniques have been successful in demonstrating affmity maturation in response to a variety of antigens: RV and RV vaccine (Thomas and Morgan-Capner, 1988; Mauracher et al., 1992a), Toxoplasmosis (}{edman et al., 1989), Respiratory syncytial virus (Meurman et al., 1992) and Varicella zoster virus (Tilley and Junker, 1992). It is still argued whether the chaotropic elution ELISA indirectly measures affinity or the avidity of immunoglobulins. In the titles of papers and abstracts which use this method both terms are used, at times -32-  interchangeably.  As chaotropic agents most likely directly interfere with the epitope-antibody binding site it  would seem reasonable to suggest that this method measures antibody affinity. To resolve this issue the interaction of low and high affinity anti-dinitrophenol (DNP) monoclonal antibodies with variably substituted DNP-albumin could be investigated by chaotropic elution ELJSA.  If this assay can selectively differentiate  between high and low affmity antibodies, independent of DNP concentration on the albumin molecules, then it could become possible to conclude that the chaotropic elution assay indirectly allows for the assessment of antibody affinity. ELISA In the ELISA technique, affinity of a serum can be estimated by the slope of the titration curve, especially if non-repeating antigens are used (Steward and Lew, 1985). Low affinity monoclonal antibodies need to be present in far higher concentration to produce similar absorbance values as high affmity antibody.  Similar  observations have also been made in inhibition ELISA, where low affinity monoclonal antibodies require higher concentrations of liquid phase inhibitor to decrease solid phase binding (Rath et al., 1988). This principle was adapted in a One-Well Inhibition ELISA where a constant amount of antibody was added to wells with constant amounts of coated antigen. The amount of liquid phase inhibitor was variable and the amount of inhibitor required to bring about a 50% decrease in absorbance (I) was found to correlate well with the average affinity of sera (Steward et al., 1991). These authors used polyclonal sera directed to an 18 an peptide representing the neutralizing domain of the foot-and-mouth disease virus and were able to verify the affinity values obtained by the inhibition ELISA with those obtained by the ammonium sulphate precipitation technique.  1.5. Thesis Rationale and Thesis Obiectives Rubella virus is a pathogen which provides the rare opportunity to study the influence of intrauterine viral infections on human immunity. The study of CRS patients as a model for the induction of immunological tolerance is of advantage, as RV biology, pathology and molecular genetics are well established and clinical specimens of both pre- and postnatally infected individuals were readily available in our laboratory. This thesis tested the hypothesis that intrauterine infection of the human fetus in the first trimester of gestation with Rubella virus leads to development of immunological tolerance. -33-  A study of the tolerization of the human  immune system with RV will help in the understanding which mechanisms for immunological self/non-self recognition are operational in humans.  Furthermore, the induction of immunological tolerance may be an  important mechanism for viral persistence.  Persistence of RV in CRS patients has been well documented.  However, RV persistence may also play a role in the pathogenesis of chronic inflammatory disorders such as RAA. The identification of immune responses which may be inefficient in clearing viral infections may therefore be of relevance in both the congenitally-infected patient and patients developing chronic rubella associated arthritis. The objectives of the thesis are discussed in point form below and also reflect the titles under which the individual sections of the results and discussion chapter are presented.  The development of reagents and  techniques, needed for the comprehensive evaluation of RV-specific humoral responses, were initially established. Once these assays had been established it was possible to first defme the normal adult response to RV infection followed by the analysis of antibody responses to RV in CRS patients.  I)  Preparations and Characterization of RV for Use in Solid Phase Immimoassays. In order to compare  the antibody responses of CRS and control patients, new virus specific immunoassays had to be developed and standardized. It was therefore necessary to produce a large and uniform preparation of RV which would provide a large enough stock to be used for all further experiments.  This virus stock was defined for its RV titre,  antigenicity and purity. Whole RV ELISA methods were optimalized using this preparation and this RV preparation was also used in the evaluation of suitable methodologies of anti-RV IgG affinity assays.  II)  RV Protein Specific Immunoassays.  The objective of the methodological development was to have  techniques available which could measure the following parameters of the RV-specific antibody response: 1) Quantitation of IgG to whole RV and RV proteins El, E2 and C, 2) Measurement of relative affinities of IgG directed to whole RV and RV proteins El, P2 and C, 3) Measurement of biological activity of sera by HAl assay and 4) Determination of specificity of antibodies to linear and topographic epitopes of the viral proteins. The development of new techniques was necessary in order to initiate the comprehensive study of patients infected with RV in the pre- and postnatal period. These included methods for evaluating the specificity and affinity of antibodies to individual RV proteins.  Immunoprecipitation, Western blotting and ELISA using -34-  separated RV proteins were evaluated as techniques suitable for the detection of RV protein-specific immunoglobulins and needed to be capable of measuring both antibody quantity and affinity. It had been reported that both El and E2 proteins contained mainly conformationally dependent epitopes and it was necessary to show that purified RV antigens retained their native structure, before using these proteins in comparative immunoassays.  III) RV-SpecifIc Antibody Responses in Postnatally Infected Patients. In order to determine if a state of RV specific tolerance existed in the CRS patients it was first necessary to characterize the humoral immune response of healthy patients to RV. Although antibody responses to RV in adults have been evaluated for more than four decades a comprehensive study of biological activity (HAl), whole virus and virus protein specificity, antibody affmity and linear/topographic epitope specificity of RV-specific antibody responses has never been undertaken. Once the RV-specific antibody response had been defined in healthy responders it was possible to analyze the responses of CRS patients.  Differences in the responses of these patient groups may reveal defects in the  immune repertoire which would suggest a state of immunological tolerance in CRS patients.  IV) RV-SpecificAntibodyResponses in Congenitally Infected Patients. To test the hypothesis that intrauterine exposure to RV had established immunological tolerance, the RV-specific antibody responses of CRS patients had to be analyzed and compared to those of the control population. If differences in the humoral response were detected, a model of tolerance could be proposed which could accommodate serological findings and current data on cellular responses in this patient group.  -35-  2. MATERIALS AN]) METHODS  2.1. Virus Preparation. Titration and Concentration 2.1. 1.Bulk Preparation of Rubella Virus (M-33 Strain) Rubella Virus, M-33 strain (VR-3 15, ATCC, Rockville MA, USA) (Parkman et al. 1962) was grown by infecting 80% confluent monolayers of Vero Cells (CRL 1586, ATCC) as follows. Vero cells were grown in complete MEM containing 1% Glutamine (30 mg/nil), 100 units/mI and 0.1 mg/nil of Penicillin and Streptomycin respectively to prevent bacterial contamination (all from Gibco, Grand Island NY, USA), 20mM Hepes (Sigma, St. Louis MO, USA) and 2% fetal calf serum (1-lyClone Laboratories, Logan UT, USA) using T-175 Falcon tissue culture flasks (Becton-Dickinson, Lincoln Park NJ, USA) at 36°C in 5% CO . Supernatant was aspirated 2 and cells were washed once with 10 ml PBS at 36°C. RV (at an m.o.i. of 10) was then added in 30 nil of medium without fetal calf serum. Virus was left to absorb at 36#C for 2 hours. 120 ml of complete medium was then added and cells were returned to the incubator for 3 days. The primary virus harvest at three days following infection was used as the starting material for further viral purification. The technical help and the many innovations which Ms. S.Farmer introduced to the growing of RV is acknowleged and greatly appreciated.  2.1.2.Rubella Virus Titration on RK-13 Cells RK-13 cells (CCL 37, ATCC) were seeded into 96-well Falcon plates, at 0.2 ml per well, at a concentration of 20,000cells/ml, using M199 Medium, 1% penicillinlstreptomycin and 2% fetal calf serum (suppliers as above). At 48 hours, medium was removed and the wells were washed with warm PBS. Tissue culture supematants containing RV were diluted in 10 fold dilution series and 50 11 loaded in triplicate onto the confluent RK-13 cells followed by a one hour incubation 36°C. Following virus absorbtion, 150 11 of complete medium was added to each well and infectous foci were scored after 3-4 days.  2.1.3.Concentration of Rubella Virus Tissue Culture Supernatant Tissue culture supernatant was clarified by centrifugation on a Silencer H-103 N centrifuge at 3000 rpm (G av= 1000) for 15 minutes. Pooled preparations were then brought to 0.5 M NaC1 and 10% w/v polyethylene  -36-  glycol 20,000 (BDH, Poole, England) while stirring slowly at 4°C for 2 hours.  The suspension was then  precipitated by centrifugation in a Sorvall GSA rotor in a RC-5B Sorvall Centrifuge (DuPont & Co, Newton CT, USA) at 8000 rpm (G av 7000) for thirty minutes. The virus pellet was resuspended in TE buffer (10 mM Tris, 5 mM EDTA, pH 7.4)  2.2. Preparative SDS-PAGE 2.2. 1.Electrophoresis RV proteins were separated on 1.5mm thick slab gels using the Protean II mini-gel system (BioRad, Richmond CA, USA).  The discontinuous buffer system according to Laemmli (L.aemmli, 1970) was used.  The  polyacrylamide separating gels (10 %I0.25 % Acrylamide/Bis, 0.375M Tris-HC1, 0.05% Ammonium Persulphate, 0.05% Temed, pH 8.8) were overlaid with stacking gels (4 %I0. 1% Acrylamide/Bis, 0.125 M Tris-HC1, .05% Ammonium Persulphate, .05% Temed, pH 6.8) and a custom-made preparative teflon comb used for well formation.  In order to counteract “smiling” of the dye front, the comb was cut with the reference wells  exceeding the preparative well by 1.5mm in length. In all cases, gels were poured and left to polymerize over night in the dark at 4°C. Sample buffer for preparative PAGE was devoid of reducing agents, contained less SDS than commonly used and was prepared monthly as double strength stock (0.25 M Tris-HC1, pH 6.8,0.5% SDS, 20% Glycerol, .02% Bromophenol Blue). 50 11 of RV preparation was diluted in 300 11 dH O followed by the addition of 350 11 2 of Sample buffer. Tubes were incubated in a 60C waterbath for 3 minutes and loaded into the preparative wells. The reference wells received 50 11 of pre-stained molecular markers (Sigma, St. Louis MO, USA). Electrophoresis was performed at 40 niAmps per gel, constant current, until the dye-front reached the bottom of the slab. The location of RV protein bands in the slab gel was detected by excising a thin lane of gel containing half the lane of pre-stained MW markers and the edge of the preparative lane containing RV, followed by transfer to nitrocellulose and immunoblotting.  The nitrocellulose strips were developed using a  human anti-RV serum containing IgG to El, E2 and C (See immunoblotting section for methods). These bands were then used as a guide in the excision of gel strips containing antigen.  -37-  2.2.2.Electroelution Gel strips were cut into 1 cm sections and electroeluted using a Model 422 electroelution attachment of the BioRad Protean II system. Elution buffer (25 mM Tris base, 192 mlvi Glycine, 0.1 % SDS) was filtered through a 0.22 im filter. Electroeluates of each protein were collected, frozen at -70°C until enough of each protein was prepared to provide sufficient material for completing the anticipated patient study. Each preparation was assayed by immunoblot and the preparations of each viral protein, showing no contamination with other RV structural proteins, were pooled. The C antigen pool showed residual contamination with El and was further purified using Triton X-l 14 extraction (see below).  2.2.3.Detergent Extraction Capsid preparations were treated with KCI for SDS removal as follows. 100 11 of 0.4 M KC1 were added to 1 ml of C preparation at room temperature.  The preparation of the C antigen pool was then put onto ice and  precipitation of the potassium lauryl sulphate salt allowed to equilibrate for 30 minutes.  Tubes were then  centrifuged for 1 minute at 10,000 rpm on a bench-top centrifuge and the SDS-free C preparation transferred to new tubes and left on ice. Triton X-114 (Sigma, St. Louis MO, USA) was added to a final concentration of 0.5% and the tubes rotated at 4 #C for 1 hour and then brought to room temperature.  Triton X-l 14 reaches  its clouding point at 17°C, above which it can be removed from an aqueous solution by centrifugation. The detergent free C preparation was then aliquoted and stored at -70 °C.  2.2.4.Yield Determination of Electroelution 5]methionine labelled RV (see below) was loaded into two wells of reducing 9% SDS-PAGE slab gels with 35 L-[ each well containing antigen being flanked by pre-stained MW markers. Gels were run until the BPB dye front reached the bottom of the gel. The region of the lanes containing labelled virus located between the MW the 58- and 26 kD markers was excised, fragmented and added to 2 ml scintillant (ScintiVerse E, Fisher Scientific, Fair Lawn, NJ) with the addition of0.4 nil of a 1% SDS solution. The other lane was electroeluted as described above and the total volume of 0.4 ml of electroeluate was added to an additional scintillation vial. Vials were kept ovemight, in the dark, and counted the following day on a Beckman LS6800 Scintillation Counter. Yields were determined by comparing the counts obtained from the electroeluate and the fragmented gel strip. -38-  2.3. Metabolic RadioIabellin of Rubella Virus Vero cells were grown in T- 175 flasks to a confluence of 80% and infected with RV (M 33 strain) as described above. After adsorption of virus 30 ml of medium was aspirated and 150 ml of fresh complete MEM (as defined above was added.  After 24 hours the medium was changed with complete MEM lacking L-methionine.  Following a 2 hour incubation this medium was aspirated and 150 of fresh medium added (1:1 of complete MEM and MEM lacking L-methionine with the addition of 18.5 MBq of L-[ S]methionine) (Amersham, 35 Arlington Heights IL, USA). ultracentrifugation.  After a further 24 hour incubation, virus was harvested and concentrated by  Tissue culture supernatant was centrifuged at 27,000 rpm (G=9O,OOO) for 2 hours, in  polyallomer tubes, each holding 38.5 ml TCS.  Pellets were washed once with 10 ml ice cold PBS and spun for  30 minutes at 27,000rpm. The supernatant was discarded and pellets taken up in 385 11 TBS.  2.4. Serum Separation Separation of IgG, IgA and 1gM from human serum was performed using anion exchange chromatography. Sera were centrifuged on a bench-top centrifuge for 10 minutes at 10’OOO rpm to remove particulate material and chylomicrons. Samples were then diluted 1:2 in running buffer and filtered through 0.22 im Millipore filters. Solutions used in chromatography were as follows: Running buffer  (Buffer A):  20 mM Tris-HC1, pH 8.3,5 mM NaN 3  High salt buffer (Buffer B):  20 mM Tris-HC1, pH 7.1,500 mM NaC1, 5 mM NaN 3  Column rinse buffer 1  200 mM NaCI, 1% Triton X-100  Column rinse buffer 2  1 M Acetic Acid  Dialysis buffer  50 mM Ammonium acetate, 10 mlvi NaCl, pH 7.2  All buffers were filtered through 0.22 im Millex-GV filters (Millipore, Bedford MA, USA) and stored in the cold room.  As chromatography was performed at room temperature, buffers were equilibrated to abient  temperature and degassed before use. Separation was performed on a Mono  Q HR5/5 (Pharmacia, Uppsala,  Sweden) column attached to a LKB 215OHPLC system (LKB, Bromma, Sweden). Sample (200 1L) was loaded onto the column and proteins were eluted by a step-gradient illustrated in Figure 18. Following each increase of Buffer B, eluted proteins were pooled. The three pools collected, contained separated serum immunoglobulin classes in the following order: IgG in pool 1, IgA in pool 2 and 1gM in pool 3. Each pool was dialysed over -39-  night at 4 #C and lyophilized. Ammonium sulphate was used as salt in the dialysis buffer in order to allow the removal of this volatile salt during lyophilization. Samples were taken up in 200 11 of dH O before analysis. 2  2.5. Gel Stamina 2.5.1.Colloidal Coomassie Stain Coomassie Brilliant Blue G-250 (Sigma, St.Louis, MO) was prepared according to Neuhoff et a!., (1988) as follows: 4 grams of Coomassie Brilliant Blue G-250 was dissolved in 250 ml of hot 7.5% acetic acid. Ammonium sulphate (44 grams) was added slowly and the solution cooled to 4 #C. The recrystallized dye was washed with cold 7.5% acetic acid, 18% ammonium sulphate solution and dried. The staining solution was obtained by dissolving 0.5 g of the recrystallized G-250 in 10 mIs of water which is added to 500 ml of a 0.45 M ammonium sulphate, 2% phosphoric acid solution. Befor staining, gels were fixed in 10% acetic acid, 1% trichioroacetic acid and then agitated over night in the colloidal stain working solution. Gels were briefly rinsed in 20% methanol, 10% acetic acid and dried.  2.5.2.Silver Stain Silver staining was performed with a commercially available kit (Bio-Rad, Richmond, CA). All glass plates were cleaned with Chromerge prior to gel casting and deionized water was used in all reagents.  2.6. Solid Phase Immunoassays 2.6.1.Western Blotting For gel electrophoresis separation of RV proteins for immunoblotting either reducing or non-reducing sample buffer was used. Buffers used were as follows: Sample buffer 2X (Reducing):  0.1 M Tris-HC1, pH 6.8,20% Glycerol, 0.5% SDS, 3% Mercaptoethanol, 0.01% BPB  Sample buffer 2X (Non-Reducing):  0.1 M Tris-HC1, pH 6.8,20% Glycerol, 0.5% SDS, 0.01% BPB  Stacking Gel:  4% Acrylamide, 0.1% Bis, 0. 125M Tris-HC1, pH 6.8,0.05% Temed and APS  Separating Gel:  9% Acrylamide, 0.2% Bis, 0.375M Tris-HC1, pH 8.8,0.05% Temed and -40-  APS Running buffer:  200 mM Glycine, 25 mM Tris, 0.1 % SDS  Transfer buffer:  200 mM Glycine, 25 mM Tris, 0.1 % SDS, 20 % Methanol  Immunoblot Sample Buffer:  PBS, pH 7.3,3% skim milk powder, 0.02% Tween 20  Immunoblot Conjugate Buffer:  Same as sample buffer but without Tween 20  Substrate buffer:  50 mlvi Tris-HC1, pH 9.6,140mM NaCl, 5 mlvi MgC1 , 0.3mg/mi BCIP, 2 0.3 mg/mi NBT  SDS-PAGE was performed on Mini-Protean II apparatuses gels using a preparative 0.75 mm comb to form wells. RV preparation (20 11) was diluted with 300 11 using either non-reducing or reducing sample buffer, heated at 60°C for 3 minutes and loaded. Separation was performed at 30 mAmps constant current run until the BPB dye front reached the bottom of the gel.  Proteins were then eloctrophoretically transferred to  nitrocellulose for 30 minutes at 30 mAmps constant current using Transfer buffer. Nitrocellulose sheets were then blocked in immunoblot sample buffer for one hour. Sheets were then cut into 4mm wide strips and added to test tubes containing 300 11 of the appropriate serum sample dilution in immunoblot sample buffer. Strips were incubated over night at 4°C, washed in three changes of PBS-Tween with a last wash in Conjugate buffer. Strips were added to test tubes containing 300 11 conjugate buffer with 1 ig/mi alkaline phosphatase-conjugated Goat IgG anti-Human IgG c-chain. Following a 2 hour incubation and a further wash, strips were developed in Substrate buffer. Quantitative analysis of immunoblot strips was performed by densitometry using a Model 620 Video Densitometer (BioRad, Richmond CA, USA). Using the reflectance mode, the apparatus integrated the intensity of each band and quantitated signal strength as OD x mm.  2.6.2.Inimunoprecipitation The conventional protein A-sepharose immunoprecipitation technique was used, with modifications described by others (Loo et al., 1986). Protein A-sepharose suspension (Pharmacia, Uppsala, Sweden) was washed in Sample buffer (20 mM Tris-HCI, 150 mM NaCl, 1 % TX-100, pH 7.3) and resuspended in a minimal volume of Sample buffer. Patient serum (30 11) was diluted in 200 11 of Sample buffer and added to 50 11 washed bead slurry, the resulting suspension was incubated at 4 C for 4 hours on a rotator. Beads were washed 3 times and S-methionine-labelled RV was added, in a volume of 200 Ii Lysis buffer (20 mM Tris-HCI, 150 NaC1, 1% 35 -41-  sodium dioxycholate, 1% TX—100, 0.1% SDS, 5 mM EGTA, 1 mM EDTA) and the mixture incubated over night at 4 #C. Beads were washed 4 times and then boiled for 4 minutes in reducing SDS-PAGE Sample buffer to release antibody and bound RV antigen, spun at 10,000rpm for 1 minute and loaded onto 10% SDS-PAGE gels and separated as described above.  Following electrophoresis, gels were soaked in Amplify (Amersham,  Arlington Heights, IL) for 5 minutes, dried and autoradiographed using X-OMAR, AR diagnostic film (Eastman Kodak, Rochester, NY).  2.6.3.Enzyme Linked Immunosorbent Assays (ELISA) ELISA was used to quantitate immunoglobulins directed to whole RV or separated RV structural proteins. All ELISAs were performed in 96-well, flat-bottom, microtitre plates using either: Immulon-2 (Dynatech, Chantilly, VA) or Falcon 3915 (Becton Dickinson Labware, Lincoln Park, NJ). ELISA buffers are listed below:  Coating buffers a) Carbonate (pH 9.6): b) PBS (pH 7.3):  15 mM 3 CO 35 mM NaHCO 2 Na , , 0.02% Azide, filtered. 3 40mM NaC1, 16mM 4 PO 1.5mM 4 H 2 Na , PO 3mM KC1,0.02% 2 KH , Azide  Blocking buffer (bb):  PBS, 2% Normal Goat serum, 0.02% Tween-20, filtered (0.22 im).  Sample buffer (Sb):  Same as Blocking buffer.  Heat Denatured Sample buffer (sbb):  Boil 25 ml of SB for 5 minutes, then immediately mix with 25 ml of ice cold SB and transfer onto ice (Mauracher et al., 1991b).  Wash buffer (wb):  PBS, 0.1% Normal Goat serum, 0.02% Tween-20.  Conjugate buffer (cb):  PBS, 2% Normal Goat Serum  Substrate buffer (subb):  1 M diethanolamine, 1 mM MgCl, pH 9.6,2 mg/mi p-nitrophenyl phosphate disodium Virus ELISA Purified RV was coated at a dilution of 1:3000 in PBS Coating buffer onto Immulon-2 plates by overnight incubation at 4 #C overnight using 100 11/well. Plates were warmed to RT for 1 hour, flicked and blocked with sb for one hour at RT with 200 11 buffer per well. Serum samples were diluted in sbb and serial two-fold dilutions were added in triplicate at 100 il/well. Following incubation at 36 #C for 2 hours the plated were  -42-  washed 3 times in wb then alkaline phosphatase-conjugated goat anti-human IgG( c-chain) antibody (Kirkgaard and Perry Laboratories, Gaithersburg, MD, USA), diluted to a fmal concentration of 1 ig/nil in cb, was added at 100 Ji/well. Following a further 2 hour incubation at 36 lC, the plates were washed 4 times and developed by adding 170 11/well subb, followed by incubation at 36C. Each plate contained five duplicate dilutions of a standard serum, whose concentration of anti-RV IgG was previously determined to be 943 lU/mi, by titration versus the WHO reference serum (WHO, 1971). Standart serum was diluted serially in two-fold dilution starting at 1:1000. The end-point of development was reached when the 1:1000 dilution of the standard curve reached a net absorbance of 1.0 AU at 405nm. Plates were analysed on a microplate reader (Model 3550, Bio-Rad) using a wavelength of 405 nm.  Using the five point standard curve immunoglobulins in the samples were  quantified and assigned JJJ/ml of anti-RV IgG using quadratic regression (Microplate Manager  ,  software, Bio  Rad). Protein EUSA Electrophoretically purified RV structural proteins, El, E2 and C, were coated onto Immulon-2 plates using PBS buffer at 44C overnight. Stock preparations of each protein were prepared, divided into aliquots and stored at -70 #C. Optimal coating concentrations varied for each viral protein preparation: El was coated at 1:60, E2 at 1:80 and C at 1:40 using PBS Coating buffer. ELISA was performed as described above. The standard serum (943 lU/nil) was analyzed on non-reducing immunoblot and the relative amounts of IgG directed to each viral protein were determined by densitometry. The percentages of protein specific IgGs were expressed as arbitrary units (AU) in realation to the 943 TU/tril of whole virus specific antibody. Accordingly this standard contained: 424 AU/mi anti-El, 405 AU/mi anti-E2 and 114 AU/nil anti-C.  2.7. Affinity Assays 2.7.1.Chaotropic Elution ELISA EUSA buffers and general protocol were followed as described above and has been described previously for both whole RV as well as for separated RV protein (Mauracher et al., 1992a). Serum samples were diluted in sbb and incubated in duplicate dilution series in RV antigen coated microtitre plates: one series to be exposed to Elution buffer (eb) (8 M urea, 150 mM NaC1, 20 mM Tris-HC1, 0.02% Tween-20, pH 9.6) and the control -43-  series to be exposed to wb. After a 2 hour sample incubation step at 364C, plates were washed twice with wb, then one dilution series was exposed for 4 minutes to 200 il/well of eb while the control dilution series was treated with the an equivalent amount of wb. Following elution, the plates were washed twice and routine ELISA protocol was resumed. The end-point of the assay was reached when the least dilution of the standard curve reached 1 absorbance unit. Elution ratios [ER(%)] were determined for each serum sample by selecting the dilutions which fell between absorbance units of 0.7 and 1.5 and were calculated according to the following formula: ER(%)  =  [(A405 Urea eluted A405 Background) % (A405 Control A405 Background)] x 100 -  -  2.7.2.One Well Inhibition ELISA This assay was adapted from the whole RV ELISA described above and was only employed using whole viral antigen preparations. A similar assay was previously described for use with peptide antigens (Rath et al., 1988). Microtitre plates were coated with a constant concentration of whole RV and blocked. A dilution series of liquid phase RV antigen was then added to the plates followed by the addition of a constant dilution of serum. Constant serum concentrations were determined by choosing the concentration of serum which fell into middle of the linear region of the titration curve on whole RV IgG ELISA.  The wells which received the highest  concentration of liquid phase antigen will result in the lowest absorbance reading, following the development of the plates, as antibody binding to the solid phase will have been prevented by the competing liquid phase antigen.  The concentration of liquid phase antigen needed to bring about 50% inhibition of antibody binding  (‘.s was related to antibody affinity, with low affmity antibodies requiring more liquid phase antigen to bring about 10.5 (Andersson, 1970). Whole RV was coated onto Falcon plates in carbonate coating buffer at a dilution of 1:3000 and sera titred out from dilutions of 1:10 to 1 :50,000in sb. A dilution which yielded 0.5 absorbance units after one hour incubation was chosen as the serum concentration to be used in the inhibition assay. In case of FPLC purified IgG fractions, the lyophilizate was taken up in 100 11 sb and then further diluted as serum samples. RV preparation was diluted from 1:5 through 1:5000 in S nil test tubes using ELISA sb, the final dilution contained no virus and was taken as the infmitely diluted virus. 50 11 of each dilution was added to each well in triplicate, using Falcon plates which had been coated overnight with whole RV (1:3000) and blocked with  -44-  sb for 1 hour at RT. A double strength dilution of the serum sample was then added to each well at 50 11/well and incubated for 2 hours at 36 41C. Routine ELISA protocol was employed in all the following steps.  The  dilution of liquid phase inhibitor to bring about a reduction of 50% net absorbance was determined by fitting a best fit line through the linear portion of the inhibition curve.  2.8. Hema1utination Inhibition Assay RV Gilchrist strain (Whittaker MA Bioproducts, Walkersville, MD, USA) was used as hemagglutinin. HA titres of this preparation were determined for each virus lot by serially diluting the HA antigen in two fold dilution starting at 1:4. The last dilution yielding complete agglutination of at total of 100 11 of a 0.125% solution of 1 day old chick eiythrocytes (PML Microbiologics, Tualatin, OR) was assigned the concentration of 1HA unit. Following HA determination the viral antigen was aliquoted and stored at -70#C. All serum samples (200 11) were pre-treated for 15 minutes with 200 11 of MnCI /heparin solution (0.5 M MnC1 2 , 2500 lU/mi Porcine 2 heparin) at 4 lC. Next, 200 11 of a 50% chicken erythrocyte (PML Microbiologics, Tualatin, OR) solution in HSAG buffer (25 mM HEPES, 140mM NaC1, 1 mM CaCl , 1% BSA, 0.25%o Gelatin, pH 6.5) was added and 2 incubated at 4 #C for 1 hour. An additional 600 11 of HSAG buffer was then added and the tubes centrifuged for 10 minutes at 1000 x G and the supernatant (now at a serum dilution of 1:8) was collected. Pre-treated serum (50 11) was serially diluted two fold in “U”-shaped polyvinylserocluster plates (Costar, Cambridge, MA) and 25 11 RV hemagglutinin, containing 4 HA units, was added to each well. After 1 hour of incubation at 4C, 50 11 of a 0.25% solution of 1 day old chick erythrocytes was added and the plates interpreted for HAl titre following 3 hours and again after an overnight incubation at 4C.  2.9. Bioloaical Function of RV Proteins  2.9.1.Solubility Shift of C Protein at Acidic pH Tissue culture supernatant (TCS) containing L-[ S]-methionine labelled RV was extracted in 1% TX-i 14 35 (Bordier, 1981) (Calbiochem, SanDiego, CA), replacing Tris-HC1 buffers withMcllvanes citrate-phosphate buffer (Mcllvane, 1962) in order to accommodate detergent extraction over pH range 4.0 to 9.0 (Figure 9). Extraction was performed with cold solutions by rotating tubes end-over-end at 4 4C for 1 hour. This solution was then overlaid onto a sucrose cushion (6% sucrose, 0.02% TX-i 14, with pH adjusted accordingly) and the tube was -45-  allowed to warm to 36 #C. Tubes were then centrifuged for 1 minute on a benchtop centrifuge, and the aqueous phase removed. Detergent pellets were washed once in Mcllvane’s buffer and then resuspended in water to the original volumes.  Add TX-114  Rotate  Transfer onto Sucrose  4°C  Centrifuge  C 0 22  Figure 9. Triton X-1 14 extraction of RV from Tissue Culture Supernatant. Ice cold TCS is made to 1% TX114, and extraction performed by end-over-end rotation at 44C. TCS is then overlaid onto a 6% sucrose cushion Following (II) and the tube warmed to RT. The TCS turns opaque as TX-i 14 reaches clouding point below the sucrose and the TCS centrifugation TX-i 14 is concentrated as a pellet (I cushion is devoid of proteins detergent. Hydrophobic proteins are in the pellet, hydrophilic remain in the TCS. .Ultracentifugation TCS containing high titres of RV was treated at either pH 7.0 or 5.0 with or without 1% TX-i 14 and with 1% SDS. The aqueous supernatants from the TX-i 14 extractions and the SDS and non-detergent treated TCS were then placed on 22% sucrose cushions and centrifuged for 3 hours at 39,000rpm in a SW41 Ti rotor (Beckman, Palo Alto, CA). The pellet was resusupended in dH O and analyzed by Western blot or by PCR. 2 Chain Reaction for Detection of RV RNA An aliquot of 50 11 from the resuspended pellets from ultracentifuged TCS was solubilized in 4 M guanidinium -46-  isothiocyanite, phenol/chiorophorm extracted and RNA precipitated in the presence of tRNA (Maniatis et a!., 1982). Complementary DNA was synthesized using M-MLV reverse transcriptase (BRL, Gaithersburg, MD). One tenth of the volume was added to the PCR mix, containing 10 c4vl of each forward and reverse primer shown below and 1 U Taq polymerase (Bio/Can Scientific, Mississauga, ONT) (Saiki et al., 1988). The primers used were both 24-mers which amplified a 287 bp fragment from the El gene of RV: 5’-TrGAACTT’CAGCCCCAAGGGGCCC-3’ and 5’-TCCCCGG1TfGCCAACGCCACTCC-3’.  Thermal cycles of 95#C for 45 sec,  6841C for 15 sec and 72#C for 30 sec, were performed on a Perkin-Elmer-Cetus thermal cycler (Mauracher et al., 1991a). The work of Mr. R.Shukin in the development and design of this protocol is acknowledged.  2.9.2.Hemagglutination Activity of Purified El Protein An aliquot of El and C protein preparation was treated with KC1 to remove SDS, as described above, and employed to competitively bind HAl antibodies (C protein was used as the negative control). seropositive human serum with an HAl titre of 1:256 was used as the source of IgG.  Pooled RV  Five rows of the  serocluster plates received 5 wells of 25 11 each of serum dilutions of 1:40, 80, 160 and 320. To each row of serum dilutions, a dilution series of El (or C) protein was added, starting with 25 11 of El preparation, 12.5, 6.25,3.1 and 0 11 of El. Each well now contained serum with El or C protein and incubated at 4 tIC for 1 hour. Each well then received 25 11 of RV containing 4 HA units and received 25 11 of a 0.5% CRBC solution after a further 1 hour incubation at 4tIC. Negative controls included cells incubated with either El or C protein but no serum, CRBC with serum and no HA antigen and CRBC in buffer only.  2.10. Patient Sera Blood samples were collected by venipuncture and sera were stored at -70 tIC until needed. CRS patients were diagnosed clinically and were shown to be positive for cord blood anti-RV 1gM and/or by RV isolation during the neonatal period. Time of intrauterine infection was established for all CRS patients on the basis of maternal history and by serological confirmation. The age of patients at the time of serum collection was 1 to 34 years. One CRS patient in the study was also diagnosed with chronic progressive rubella panencephalitis. Control patients were adults and were either mothers of CRS patients used in the study or individuals who were infected during the 1984/85 rubella epidemic in British Columbia. -47-  2.11. Statistical Methods Statistical analysis using the unpaired T-test was used for most population comparisons and distributions tested for normality using the Martinez and Iglewicz test. Level of confidence was set at a=O.05 for all analyses. Populations showing non-normal distribution were analyzed using the Mann-Whitney U test. Analysis of variance (ANOVA) was used when multiple comparisons of populations were performed. OneWay AJ’TOVA was used with level of confidence set at a=O.05. Means were separately analyzed with the Fisher’s LSD post-hoc test. Means and standard deviations (SD) were determined using descriptive statistic on the software package. All error bars in the figures of this thesis are  —  1 SD.  All statistical calculations were performed on NCSS 5.01 software (J.L.Hintze, 1987). The help of Mrs. R.Milner and Mr. M.McKinnon from the Statistical Research Support Unit, Research Centre, in the statistical design is greatly appreciated.  -48-  3. RESULTS AND DISCUSSION 3.1. Section I: Vfrus Preparation and Use in Solid Phase Immunity and Affiiilty Assays. At the start of the project preparations of RV had to be established which were uniform, well characterized and would be of sufficient quantity to provide antigen for both phases of the experimental part of the thesis: methodology development and patient studies. This was important to eliminate RV antigen preparation as a variable in the comparative studies of the immune responses of pre- and postnatally RY infected patients.  3.1. 1.Rubella Virus Preparation The RV stock, which was used for all direct antigen assays and for the preparation of purified RV proteins El, E2 and C was obtained from a pooi of 24 liters of tissue culture supernatant (TCS), and purified as one batch. The virus was obtained from ATCC (M-33 strain, ATCC VR-553) and further experiments verified the stock to contain infectious RV. This wild type RV isolate was used as antigen in all serological assays, as responses to wild type infection (both in post- and prenatally infected individuals) were analyzed. Titration The original TCS contained RV at a concentration of 2.7 x 106 pfu/ml, as determined by virus titration on RK 13 cells (Nawa, 1979). The morphology of the infectious foci (Fig. 11) on RK-13 cells were characteristic for the cytopathic effect caused by RV in this cell line (Nawa, 1979). 5—  -  -f  Figure 10. Micrograph of RV Infectious Focus on RK-13 cell Monolayers. Phase contrast microscopy (500 x) of a RK 13 cell monolayer 3 days after in vitro infection. -49-  Figure 1 1A shows whole protein stain and Western blots of the RV concentrate (lane 1), the starting TCS (lane 2) and TCS obtained from mock-infected Veto cells. The final concentrated viral preparation contained 2 x 1010 pflilml and had a 150 fold increase in antigenic activity per gram protein. Protein bands for El, E2 and C were  A Non-Reducing 12  ElEl E1E2  2  1  3  Reducing 34  MW  —  -  Cd  -*  El E2  1  4 .z  3  ElE2 E2 E2  97 -68  MW  ——97  El E2  -45 —  2  -45  C  —27  —27  B  1  2  3  45  6  7  7  8  97 68  BSA  4  45  El  E2  27  Figure 11. A) Western Blot and Colloidal Coomassie Blue Stain of RV Preparations. 0.5 11 RV concentrate (1), 10 11 starting TCS (2) or mock-infected Vero cell TCS (3). Prestained MW markers (4). MW markers are indicated on the right side of the gel. Western blots were developed with anti-RV reference standard human serum, anti-human IgG (c-chain) alkaline phosphatase-conjugate and BCIP/NBT stain. Relative positions of RV proteins are indicated on the left side. El, E2 and C refer to monomeric forms of RV structural proteins, Cd dimeric capsid. Dimeric forms of the envelope proteins are indicated as El :El (homodimer) and El :E2 (heterodimer). B) Silver Stain and Autoradiograph of concentrated RV Separated by Reducing SDS-PAGE Lanes 1 through S-methionine 4 contain a BSA titration (20,5, 1 and 0.2ng/band). RV preparations, contained 1/10 volume of 35 labelled RV, were loaded in doubling dilutions in lane 7,6 and 5 (10,5 and 2.5 11 of RV concentrate respectively). Lane 8 contains MW markers. The corresponding autoradiograph of lane 7 is shown to the right with relative in the gel indicate possible El and C bands. positions of RV proteins indicated. Arrow heads .  -50-  not clearly identifiable by either colloidal Coomassie or silver staining. The detection limit for protein in the silver stained gel in figure 1 lB lies between 1  -  .2 ng BSA per band.  If the faint bands () in lane 5 are  interpreted to represent the El and C protein, then the final protein concentration of RV antigen can be estimated to lie in the order of 1 ig/mi.  The relative molecular masses of the RV structural proteins, as  determined by reducing SDS-PAGE were: El (58 kD), E2 (47-42 kD) and C (33 kD). Under non-reducing conditions, the capsid protein was detectable in its dimeric form (Cd) at 68 kD and El travelled at a lower apparent molecular mass of 55 kD, in comparison to its position on gels run under reducing conditions where its apparent mass was estimated to be at 58 kD (Baron et al., 1991, Mauracher et al., 1992a).  3.1.2Whole Rubella Virus ELISA Although basic conditions for whole RV ELTSA had been established (Bidwell et al., 1977; Tingle et al., 1989) and numerous commercial suppliers have marketed reliable anti-RV IgG ELISAs (Dimech et al., 1992), a series of modifications to the whole RV ELISA protocol were developed.  These improved both sensitivity and  specificity of the RV ELISA. Conditions and Antigen Treatment Whole RV was coated onto 96-well ELISA plates (Falcon 3915 or Immulon-2) diluted at 1:100  -  1:25 ,000in  PBS (pH 7.3) followed by blocking in NGS-containing Sample buffer. The addition of Triton X-l00 to the whole RV preparation improved signal strength (Fig. 12). Whole RV was made to a fmal concentration of either 2%, 0.5% or 0.2% TX-100 and compared to RV, without any detergent added, in a titration of antigen using constant concentration of detection antibody. Whole RV, in the absence of detergent, showed a shallow titration curve with a saturated region at dilutions stronger that 1:500. A marked increase in signal was observed between dilutions of 1:200 and 1:6400 for viral preparations which had been treated with detergent. A strong prozone was observed in the preparations containing 0.5% and 2% TX-100 suggesting that detergent may compete for hydrophobic binding sites on the polystyrene plastic surface. The increase of the signal in the detergent-treated virus may result from disaggregation of the virus or to the disruption of the viral envelope leading to increased exposure of the capsid as well as E2 protein.  -51-  2.0-  0-0 No n—s 0.2% h--a 0.5% a--. 2%  —  /  TX—laO TX—laO TX—100  /  1.5-  1.0-  TX—100  e-L /  /  A  a 1.  0 C’)  /  0.5-  0.0— 100  I  I  200  400  800  I  I  1600  3200  I  6400  —  12800 25600  Reciprocal Dilution of Whole Rubella Virus  Figure 12. Titration of RV preparations Treated with Triton X-100. Virus was brought to a final concentration of 0.2% (1), 0.5% () or 2% () TX-100 using a 20% stock solution. As a control an equivalent amount of PBS was added to the preparation which received no detergent (0). Denatured Sample Buffer During the course of these investigations it was frequently observed that sera from certain patients had high backgrounds on RV IgG ELISA. Such high background could result in false positive RV serology. Increased backgrounds were especially common in infants (Dr. L.A.Mitchell, Dept. of Pathology, UBC, personal communication), who had RV vaccine as well as several other vaccinations during the same period. It was observed that much of the non-specific background in such patients could be removed by pre-adsorption of serum on NGS blocked microtitre plates (Fig. 13). This additional step was time consuming and costly and it  -52-  was found that it could be replaced by including heat-denatured protein in the Sample buffer (Fig. 13) (Mauracher et al., 1991b). Briefly, SB was boiled for 5 mm. and then mixed with an equal volume of ice cold SB. The resulting buffer, termed “SBb. for boiled sample buffer, contained heat denatured and aggregated proteins which remained in suspension. Table 5 shows the quantitation of serum IgG directed to whole RV as well as the individual structural proteins: El, E2 and C. Two vaccinees, 2 and 3 showed no difference in their quantitative level of anti-RV IgG, expressed in TU/ml when either SB or SBb was used as a diluent. These observations indicate that sample buffer containing the heat denatured protein did not reduce background levels by masking RV-specific signals. A value of 20 lU/mi of anti-RV IgG was determined in the pre-immunization serum of patient 1 using SB. While this value was in the range of positive immunity (ie., equivalent dilution prepared in SBb reduced RV-specific IgG levels to 0 lU/mi.  15 lU/nil), an  RV seronegativity was also  confirmed by determination of a HAl titre of < 1:8. These observations suggest that SBb can aid in the  reduction of non-specific binding, without interfering with RV antigen-specific binding.  Preabsorbed —e Neg. Serum  Regular SB n-n BoHed SB 0—4 Neg. Serum  0.3-  G—E)  E  C  ---  IC)  °  02-  0 C) C 0 -D  L.  0 Cl) -Q  0.00  I  160  260  360  400  0  Reciprocal Serum Dilution  100  200  300  400  Reciprocal Serum Dilution  Figure 13. Determination of RV-Specific IgG by ELISA. A single patient’s serum sample, taken prior to RA 27/3 immunization, was serially diluted in regular sample buffer (0) or in SBb (1). An equivalent dilution series in regular sample buffer was preadsorbed on NGS-blocked micropates (s). Negative control (o). -53-  Proteins may be structurally altered by binding to plastic surfaces (Van Regemnortel, 1990). Also, it has been shown that immunoglobulins which do not bind to liquid phase antigen will recognize new epitopes (cryptotopes) revealed by structural deformation of protein occurring during solid phase binding (Friguet et al., 1984). Similarly, new epitopes could potentially be created by the juxtapositioning of monomeric proteins onto solid phase (neotopes).  It is possible that such cryptotopes and neotopes are exposed or created during heat  denaturation and aggregation respectively. The decrease of non-specific binding, described above could be attributed to the competitive removal of antibodies reactive to denatured NGS constituents in the liquid phase of the diluent. Immunoglobulins reactive to such common blocking agents as BSA and normal animal sera have been previously reported (Levinson and Goldman, 1988) and are thought to have arisen as a result of activation of the immune system with dietary antigens or by vaccine contaminants. Vaccinee  Weeks postvaccine  lgG rubella ELISA (lU/mI) Boiled sample buffer  Regular sample buffer El  E2  C  Whole RV  El  E2  C  Whole RV  1  0 6 54  4 21 51  5 64 49  8 28 16  20 62 156  0 7 20  0 21 20  0 10 4  0 24 52  2  1 6 64  0 10 103  0 8 70  1 4 10  0 28 306  0 12 108  0 8 75  0 6 12  0 30 321  3  0 6 54  0 19 58  0 26 46  1 4 6  0 60 180  0 20 60  0 22 42  1 3 7  0 64 198  Table 5. Rubella IgG ELISA Responses to Whole RV and RV Structural Proteins. The level of IgG directed to viral antigen was determined using the two sample buffers SB and SBb in three individuals at sequential time intervals following immunization with RA 27/3 Rubella vaccine. (From: Mauracher et al., 1991b)  3. 1.3.Affinity Assays For the measurement of affinity, the ideal experimental situation requires that both antibody and antigen be present in pure form (Steward and Steensgaard, 1983), such as monoclonal antibody reacting with haptenic antigen. Well established techniques, such as equilibrium dialysis (Eisen and Karush, 1949) and ammonium sulphate precipitation (Steward and Petty, 1972) have found application in such antibody-antigen studies. However if the affinity of IgG responses from sera are to be estimated in an antigen system which contains multiple proteins, then the traditional techniques for affinity measurement are not applicable (Hedman et al., -54-  1988). Two techniques which had been previously reported for their ability to measure average affinities of IgG from serum (Steward et al., 1991; Inouye et al., 1984) were investigated for their applicability in the RV antigen system.  Experiments were performed to establish the ability of these techniques to demonstrate affinity  maturation of anti-RV serum IgG following primary RV infection in humans. Inhibition ELISA This ability of this technique to measure relative affinities of sera to RV was evaluated using patient sera collected at sequential time intervals after onset of natural Rubella infection to determine if the affinity maturation of the IgG response in the first 3 months following infection could be demonstrated.  Prior to these  experiments it was established that the concentration of liquid phase antigen required to bring about a 50% reduction of the signal obtained with no competitor antigen (l) was independent of the constant antibody  .—. 12 Days Post Symptom •.- 8 Months Post Symptom  a)  0 C Q L.  0 cn  0  10’OOO  100  1000  10  Inhibitor Dilution (1/x)  Figure 14: Determination of ‘05 in Acute and Convalescent Sera. Serum samples were obtained at 12 days (acute) and at 8 months (convalescent) following diagnosis of Rubella rash. The specific absorbance of the serum in the absence of competing antigen is shown on the left side of the y-axis (0.49 and 0.67 for the convalescent and acute sample respectively), with the 50% absorbance values indicated on the right hand side of the axis. The ‘0.5 values were extrapolated from the best fit line through the linear region of the titration curve and are indicated as underlined values on top of the x-axis. -55-  concentration chosen, as long these antibody concentrations were on the linear region of the  RV ELISA  titration curve. For all sera tested the 10.5 were identical at serum dilutions of 1:250 to 1:1000. It was however required to plot RV EUSA titration curves for each serum in order to establish the linear range dilutions. Figure 14 illustrates the determination of the Io for an acute (12 days) and convalescent (8 months) serum sample following the onset of a rubelliform rash in an adult patient. The ‘0.5 values were determined in three separate experiments for paired (acute and convalescent) serum samples from 3 patients. The results shown in Figure 15 reveal that this technique could demonstrate affinity maturation of anti-RV IgG in the months following Rubella, although patient numbers were not of sufficient size to determine statistical significance.  1000-  1-00  -  10 812  Days  18  6  8  Months  Time foflowing Onset of Rubella Rash  05 were measured I Figure 15. Affinity Maturation of Anti-RV IgG Measured by Inhibition EUSA. The values in three individual patients (0, 1) at 8, 12 and 18 days or 6 and 8 months following the onset of a rubelliform SD. rash. Each patient sample was tested three times. Error bars indicate ,  -56- Urea Elution ELISA Low affinity antibodies are easily eluted from their antigen by briefly exposing the antibody-antigen complex to denaturants such as 8 M urea, whereas high affmity antibodies are resistant to such treatment. This differential sensitivity of low and high affmity antibodies to chaotropic agents provides the basis for the 8 M urea elution EUSA. It was initially established that a 4 minute exposure of antibodies complexed to solid phase-bound RV to 8 M urea, provided the optimal conditions for differentiating the low affinity anti-RV IgG in acute sera from high affinity anti-RV IgG in convalescent sera (Fig. 16). The antigenicity of solid phase bound RV was not adversely affected by exposure to 8 M urea, nor was antigen eluted from the microtitre plate by such treatment. This was established by exposing RV-coated and blocked niicrotitre plates to a 4 minute wash of 8 M urea prior to the addition of a 1:1000 dilution of anti-RV standard serum (Fig. 17). This simple technique allowed the rapid determination of relative affinities of anti-RV IgG in human sera and was capable of demonstrating the affinity maturation of these antibodies in the months following Rubella infection. 065.5%  16 1  6 months post rash (PBS) 6 months post rash (Urea) 2 weeks post rash (PBS) o-.-- 2 weeks post rash (Urea) °—° 0- - -o  57.9 % 4.8 %  -  a.  • 52.9 X 5.1 %  0.8-  54.0 %  0.60  7.6 Z  -  <  0.40.2-  400  -  -  -i. —  800  —‘s.-.  1600  3200  6400  Dilution Factor Figure 16. 8M Urea Elution ELISA for Determination of Antibody Affinity. Paired serum samples from a patient taken at 2 weeks () and 6 months (I following the onset of a Rubella rash. The dilution series washed in WB only are indicated by solid lines, the 8 M urea washed dilution series is indicated as a dashed line. Above each point the calculated elution ratio [ER( %)] is given, showing that ER( %) remains relatively constant over a wide absorbance range. -57-  e—o PBS WASH 0-,  8M UREA  1•0 0 0 0 C  0.5-  0.01000  2000  4000  8000  16000 32000  Dilution of Whole RV Preparation (1/x) Figure 17. Treatment of Solid Phase Bound Rubella Virus with 8M Urea. Whole RV was coated in two-fold dilutions onto microtitre plates and then either washed for 4 minutes with 8 M urea wb (o) or with wb (Iv, in order to determine whether 8M urea affected the antigenicity of RV. Standard ELISA protocol was then followed using a 1:1000 constant dilution of anti-RV standard serum to detect RV antigen. of IgG Fractions from Human Sera Both techniques described above were capable of measuring low affinity anti-RV responses in the first weeks following Rubella. Although the primary response is initially characterized by a low affinity IgG response, it is also typical to detect high levels of antigen-specific 1gM in the first weeks following infection. Hence, it was important to determine if the presence of 1gM in RV antisera influenced affinity measurements of anti-RV IgG. For that purpose, lgG fractions were prepared by separating this immunoglobulin class from IgA and 1gM using FPLC anion-exchange chromatography. The FPLC system allowed rapid and repetitive runs of 100 11 serum. The elution profile of33%, 48% and 70% buffer B, allowed the best separation of three immunoglobulin classes G, A and M (Fig. 18A). The overall yield of IgG was 60%, with the yield of IgG in the purified IgG pooi being 50%. Losses of protein may have occurred during filtration of sera, losses of protein in peak trails which were  not included in the pooled fractions or during dialysis. The dialysed Ig fractions were lyophilized and stored at  -  70#C. IgG fractions and sera were compared for three patients, each containing an acute and a convalescent sample, using the One-Well Elution ELISA and the 8 M Urea Elution EUSA. It was shown that the presence of RV specific 1gM could adversely influence the measurement of IgG affinity when using the One-Well -58-  A  1 .0  100  E C  I1)  C”  .5 ci)  C  4-  9-  D  ci) C-) C 0  -c 0  (I,  -o  0  Fraction Number  B  1400EZZI g G ZI g A LEn Ig M  1200-  C  0  1000-  4-,  -4-,  C ci) C) C  800-  0  (-) C D -D  0  400  0’  0 C D  E E  200  G  M  A  SER  Pooled Fractions  Figure 18. A: Elution Profile of Serum Separation. Each fraction contained 1 ml and a step gradient was run as indicated by the solid line. Eluted protein is indicated by the dashed line. Three pools were collected, each containing 5 mIs. B: Quantitation of immunoglobulins in each of the pooled fractions (G, A or M) and serum (SER). -59-  Inhibition ELISA (Table 6). It is likely that RV-specific 1gM binds liquid phase antigen resulting in the increased requirement of liquid phase antigen to compete out IgG binding. This results in a shift to the right of the inhibition titration curve and in an overestimation of the  In contrast, the affinity measurements using  ER(%)  150 Serum  IgG Fraction  Serum  IgG Fraction  Patient 1 acute convalescent  8 53  5 48  7 390  50 400  Patient 2 acute convalescent  1 52  2 47  20 398  63 410  Patient 3 acute convalescent  12 70  18 71  50 450  60 450  Table 6. Relative Affinity of Anti-RV IgG in Serum or Purified IgG. Serum samples from three patients were collected during the acute and convalescent period following Rubella. IgG fraction, prepared by ion-exchange chromatography, and whole sera were compared for their relative anti-RV IgG affinity measured by using both 8M Urea Elution ELISA (ER(%)) and the One-Well Inhibition EUSA  the Elution EUSA were not affected by the presence of RV-specific 1gM. The overall amount of RV-specific IgG binding to the solid surface may well be decreased by competing 1gM, this however should not affect the ratio of low and high affinity antibodies, as measurements of ER(%) were performed in antigen excess (ie: in the linear region of the titration curve).  Comparison of the One-Well-Inhibition ELISA and Elution EUSA The inhibition EUSA showed two main disadvantages over the Elution EUSA: 1) The presence of anti-RV 1gM interfered with the determination of affinity values when using the inhibition assay. The valence of 1gM may give this molecule an advantage over IgG for binding antigen in the fluid phase.  The reason why this interference of 1gM with affinity measurement of IgG is observed remains  somewhat of a puzzle.  Other groups using a similar assay with peptide antigen, have not observed any  -60-  difficulty with 1gM in acute sera (Dr. M.W.Steward. personal communication). 2) The amount of antigen required for the liquid phase inhibitor exceeds the requirement of antigen by the Elution ELISA by 100 to 1000 fold. This would have provided a significant obstacle for the use of purified RV proteins. Based on the above describe observations, it was decided to proceed with the Elution ELISA as the assay of choice, which would be employed to study the relative affinities of sera to whole virus as well as to the separated RV structural proteins El, E2 and C.  -61-  3.2. Section II: Protein Specific Immunity: Antibody Quantitation and Affinity to Separated RV El, E2 and C. Rubella virus contains three structural proteins which are immunogenic in humans. However it is the El-specific antibodies which have hemagglutination inhibition activity and the majority of the in vitro neutralization activities. In mouse monoclonal antibody studies 4 El-specific epitopes have been defined which have both in vitro neutralizing (in absence of complement)- and hemagglutination inhibition activity. A further 2 epitopes have been described to have hemagglutination inhibition activity only. On the other hand only 1 epitope has been defmed on E2 which can neutralize RV replication in vitro (Green and Dorsett, 1986). It is therefore believed that the response to the envelope protein El is paramount in protective humoral immunity. As whole RV assays, such as those described above, do not differentiate between protein specific responses, it is necessary to develop protein specific immunoassays to quantitate the antibody responses directed to individual viral proteins. Assays such as immunoprecipitation (Katow and Sugiura, 1985), purified RV protein ELISA (Mauracher et a!., 1992a) and Western blotting (Zhang et a!., 1992) have allowed the investigation of protein-specific immune responses following Rubella infections. This section will discuss the development and some applications of the latter two techniques referenced above. 3.2.1.Western Blotting  3.2.1.lReducing versus Non-Reducing Conditions Figure 19 illustrates a Western blot and autoradiograph of gel separating viral proteins under reducing ME) in the left lane and non-reducing conditions  ( + 2-  (- 2-ME) in the right lane. This SDS-PAGE separation was  unorthodox in that both wells were positioned side by side to each other and the diffusing of reducing agent during the stacking run into the right lane, which contains non-reducing sample buffer, illustrates the decrease of both El and E2 antigenicity in the presence of reducing agents.  Capsid, under non-reducing conditions  travels as a dimer (Cd). Its antigenicity is not stronger in the absence of reducing agents, and Cd is poorly visible in the Western blot in Figure 19. The autoradiograph of the Western blot shows that the increase in signal strength observed in the immunoblots is due solely to antigenicity and not to quantity differences. The apparent molecular mass of El was observed to increase by about 4 kD upon reduction. This suggests that native El is held in a compact tertiary conformation. El contains an average of 5 mole % cysteine residues (Clarke et a!.,  -62-  1987) which likely contribute to the highly folded conformation of this protein. Although E2 contains a similar distribution of cysteines, its apparent MW is not affected by reducing conditions. The C protein travelled as a dimer under non-reducing conditions, and the transition from the 66 kD dimer to the 33 kD monomer can seen in the autoradiograph in Figure 20A. As the addition of reducing agents selectively affected the antigenicity  B  Western Blot  A  2-ME +  Autoradiograph 2-ME +  —  D$  D <Cd  El =-  E1  <El  E1  E2i  E2I  <E2  ct.  Figure 19. Western Blot and Autoradiograph of Reduced and Non-Reduced Preparations of RV. A: Western blot developed with polyclonal human anti-RV serum as primary antibody. Both lanes contain equal amounts S]methionine labelled RV, run in the presence (+ 2-ME) or absence (- 2-ME) of 2-mercaptoethanol. 35 of L-[ Reduced ( ) and non-reduced ( ) viral structural proteins are indicated at the margin: (D) homodimeric and heterodimeric forms of El and E2, (Cd) capsid dimer, as well as the monomeric forms of El, E2 and C. B: Autoradiograph of the Western blot illustrated in A. (Mauracher et al., 1992a)  of El and to a lesser extent that of E2. Western blot protocols were put into practice which excluded the use of reducing agents.  It became apparent that non-reducing blots were about between 1  -  2 log units more  sensitive than those performed under reducing conditions and evidence by others indicates that the non-reducing blots may shows equal sensitivity as RV IgG ELISA (Dr.L.A.Mitchell and Ms.M.Ho, Dept. of Pathology and Pediatrics, UBC, personal communication).  Before the development of the immunoblotting technique for the  detection of RV specific IgG, immunoprecipitation of radiolabelled RV was the assay of choice for the detection -63-  of protein-specific IgG, and had been used in two studies to determine the protein specific responses in rubella responders (Katow and Sugiura, 1985; de Mazancourt et al, 1986). If the immunoblot technique was to replace the accepted methodology for the measurement of protein specific responses it had to be demonstrated that inmiunoblot provided similar response profiles as observed with immunoprecipitation.  3.2. 1.2.Comparison of Immunoprecipitation and Western Blot under Non-Reducing Conditions Six patient sera were compared for their RV protein-specific IgG responses using immunoprecipitation and imniunoblot assays performed under non-reducing conditions (Fig. 20). The results show that immunoprecipitat ion overestimates the amount of anti-E2 IgG present in serum. Lane 2 contains a serum with no E2 specific S]-methionine 35 antibody as determined by the immunoblot technique. This may be due to two factors. If L-[ is used as a metabolic label then differential distributions of this amino acid in the RV proteins has to be taken into account. El contains only 3 methionine residues whereas E2 and C contain 7 and 6 respectively (Clarke  Immunoblot 1  2  3  4  Immunoprecipitation 5  6  1  2  3  4  5  .6  Cd El  •  E2  E2  c  Figure 20. Comparison of Anti-RV IgG Western Blot and Immunoprecipitation Assays. Six serum samples from one negative control and 5 RV serology positive patients were subjected to comparative analysis by Immunoblot or Immunoprecipitation. Lane 1: negative control; lanes 2-4 responders to wild RV infection; lanes 5 and 6: CRS patients.  et al., 1987). This problem could be overcome by replacing methionine with tritiated amino acids in order to ensure equal labelling of RV proteins. The problem of co-precipitation of El and E2 was more difficult to solve. Monoclonal mouse IgGs directed against El often precipitate E2 protein (Brush, 1989) as dimerized -64-  forms of the envelope protein are resilient to dissociation even in the presence of ionic detergents.  The  discrepancy between immunoprecipitation and Western blot techniques with respect to detection of E2 signals may therefore be due to co-precipitation in the immunoprecipitation assay. The two techniques correlate well in respect to El- and C-specific signals.  -  PatIent 1  -  MW1  IgG  J 84  Patlent2  Patient 3  -.  23456789  -  101112 D  —  4E2 L  —-  1  23456789101112  r  —  —  IgA  L’_ 1  1gM  -  2  3  4  6  -  _E  :__  5  i E  789  10 1112  0 rEFrL  84 58  I  I  Figure 21. Immunoblot Analysis of RV-Specific IgG, 1gM and IgA antibodies at intervals following wild Rubella Infection. Western blots of RV proteins separated by non-reducing SDS-PAGE were developed with representative sera obtained at the following time iptervals after the diagnosis of a rubelliform rash: lanes 1, 5 and 9: 0-6 days (early acute phase); lanes 2,6 and 10:7-30 days (acute phase); lanes 3,7 and 11: 1-11 months (convalescent phase); lanes 4, 8 and 12: 1-3 years (late convalescent).  Western blotting provided a good alternative to immunoprecipitation and showed considerable advantages: a) independence from radiolabelled reagents; b) technical difficulties such as unequal protein labelling and co precipitation are avoided; c) rapid results ; d) more amenable to the processing of large samples and e) allowing the determination of RV protein-specific 1gM and IgA antibodies in addition to IgG. The immunoblot technique was sensitive enough to detect IgA and 1gM directed to the individual RV proteins (Fig. 21). An analysis of three patients, recovering from wild type RV infection shows the rapid increase of RV specific 1gM in the first months of RV infection, followed by its decline within a year following infection. Both  -65-  IgG and IgA responses are maintained in the first three years following primary rubella. The kinetics of proteinspecific responses of IgG and IgA are discussed further in section ifi. Blot AffinityAssay The principle of chaotropic elution of low affinity antibodies was applied to Western blots (Fig. 22). Patient sera were applied in duplicate to nitrocellulose strips containing blotted RV proteins, separated by reducing SDS  a  a  b  8m  12d  8m  18d  a  b  a  b  b  —  El  C  V.S.  D.H.  Figure 22. 8M Urea Elution of Western Blot. Paired serum samples from Patient A and B, taken during the acute (18 or 12 days) and convalescent phase (8 months) following natural Rubella infection. Strips labelled “a” represent nitrocellulose strips washed in PBS buffer following serum incubation. Strips “b” show blots incubated with the same sera as in lanes “a” ,but were washed for 3 minutes in 8M urea elution buffer. Position of El and C is indicated on the right side. Anti-C IgG was readily eluted by urea, both in acute and convalescent samples whereas anti-El IgG resisted urea wash in the convalescent samples, indicating a differentially higher affinity of anti-El IgG in the convalescent sample. PAGE, one strip was washed in PBS, the other to be washed for 3 minutes in elution buffer, as described above for affmity ELISA. As was observed with whole RV ELISA, signal strength was not effected when blots were washed with urea prior to the addition of antibody. The paired, acute and convalescent, serum samples obtained from patients following Rubella showed antibodies reactivity to El and C, but not to E2. The El-specific IgG matured in its affinity and became resistant to elution in the 8 month samples. In contrast the anti-C IgG was not resistant to urea elution in either the 18, 12 day or 8 month samples, suggesting that these antibodies did  -66-  not undergo affinity maturation. These results indicated that the affinities of antibodies directed to the individual viral proteins may not be equally distributed. Differential affinities of antibodies to proteins of a viral pathogen had not been reported previously and the possibility that one protein was immunodominant and was giving rise to both higher levels as well as higher affinity antibodies was investigated further (Section ifi). Most investigators which have analyzed the specificity of anti-RV antibodies have employed either immunoprecipitation (Katow and Sugiura, 1985; de Mazancourt et al., 1986) or immunoblotting under reducing conditions (Partanen et al., 1985; Cusi et a!., 1989). With the new observations reported here, concerning sensitivity and applicability of imniunoblotting performed under non-reducing conditions to the detection of Ig of all classes (Mauracher et al., 1992; Zhang et al., 1992) this technique may find application in serology during the early post-conception period to verify RV infection. The ability to determine specificity of IgA for individual RV proteins may aid in the understanding of the role of both serum and mucosal IgA in the protection against viral reinfection and provide a technique to further our understanding of the role of persisting serum IgA as an indicator of persistent RV infection (Morris et a!., 1985). Some disadvantages remain inherent to the Western blot technique. The technique is more time consuming and costlier than ELJSA and the quantitation of peaks by densitometry can introduce bias in the interpretation of protein specific signals. It became desirable to obtain RV proteins El, E2 and C in pure form and to utilize these proteins in the EUSA protocol.  3.2.2.Seyaration of RV Proteins. Their Use in Solid Phase linmunoassays Many strategies for the separation of viral proteins have been described. The properties of El, E2 and C, such as size, charge or solubility, can be used to purify these proteins. The starting material for the separation of the RV proteins was the highly concentrated virus stock described earlier in section I, which was highly antigenic and also immunogenic in animals. The goal of the protein purification strategy was not necessarily to arrive at preparations which showed high levels of purity for each protein, but to arrive at preparations which were pure in respect to other RV proteins. Hence the term “RV protein separation” will be used rather than “protein purification”. Although the envelope proteins El and E2 can be readily separated from the capsid protein using detergent -67-  extraction (Mauracher et al., 1991a) (Fig. 29), the separation of El from E2 was more difficult. Both envelope proteins have similar solubilities and have overlapping p1 ranges due to glycosylation with complex sialylated Nlinked sugars (Lundstr dm et al., 1990). More importantly they form heterodimers which are resistant to all but the most severe denaturing conditions. As described above, non-reducing immunoblots showed similar immuno reactivities as immunoprecipitations and supported the concept that SDS-PAGE under non-reducing conditions and low heat may yield separated RV structural proteins with antigenic properties similar to that of native  A  Top 1 ml Gel Silo Electro Eluate Hg. 1. Side view of Model 422 Electro-Eluter Assembly. The vertical glass tube is filled with elution buffer. The negative electrode is at the top. The positive electrode is below the membrane cap. Macromolecules are carried by the electrical current out of the gel stice. through the tnt arid into the membrane cap. The molecules are retained by a dialysis membrane which is molded into the cap.  B  Viral Protein  PRE Gel Slice  Electro Eluate  Top 1 ml  POST Gel Slice  Yield %  C  1737  1960  120  76  112  El  939  923  65  62  98  E2  2067  2025  108  72  98  Figure 23. Yield of Radiolabelled RV Proteins after Electroelution. A: Bio-Rad model 422 electro-eluter assembly. The vertical glass tube is filled with elution buffer. The negative electrode is at the top, the positive electrode is below the membrane cap. RV macromolecules are carried by the electrical current out of the gel slice, through the frit and into the membrane cap (400 11). The molecules are retained by a dialysis membrane (MW cutoff 12-15 kD) which is moulded into the cap. (From: Bio-Rad, manual 87-01330387, 1987) B: Yield of radiolabelled RV proteins by electroelution. Counts per minute (cpm) were determined in gel slices before and after electroelution (PRE and POST), “Top 1 nil” indicated a 1 ml sample of elution buffer taken from above the frit. -68-  protein. Using preparative SDS-PAGE, followed by electroelution provided a feasible strategy for RV protein separation (Mauracher et al., 1 992a). of Electroelution; Yield and Purity of RV Proteins. S]methionine labelled RV was 35 The yield of this separation technique was initially established as follows: L-[ separated on 10% acrylamide slab gels and the relative positions of the viral proteins were determined by an overnight autoradiograph of one gel lane. The RV proteins El, E2 and C were excised from the gel, with onehalf being electroeluted and the other half stored for the determination of radioactivity prior to elution. Samples of electroeluted proteins in the membrane cap, the electroeluted gel slices above the frit and elution buffer in the glass tube were solubilized in the scintillant and counted for radioactivity and was compared to the counts in the gel slices which were not electroeluted. Gel slices and electroeluate were added to I ml of scintillant and left in the dark over night. Counts per minute from pre- and post-eluted gel fragments, electroeluate and elution buffer as well as yield determinations are shown in Figure 23. The yield of protein was 90  1  2  3  -  100%. This was  4  D Cd El E2  Figure 24. Western Blot of Separated RV proteins. Preparations of capsid dimer (Cd), El and E2 proteins were run on non-reducing SDS-PAGE in lanes 1, 2 and 3 respectively. Purified whole RV was loaded in lane 4, all viral structural proteins were visualized including the dimerized species of El and E2 (D). (From Mauracher et al., 1992a) -69-  also confirmed by electroeluting known amounts of BSA and determining protein concentrations following electroelution (data not shown). The monomeric species of the RV envelope proteins, El and E2, and the C dimer differ in mass by about 10 kD each. This difference in MW is sufficient to separate these proteins on 10% SDS-PAGE slab gels. Gels were run until the 27 kD pre-stained MW marker reached the gel front, then the proteins were excised using the pre-stained MW markers as guides and electroeluted. Each electroeluate was checked for purity by Western blot and the preparations shown to be free of the other two respective viral proteins were pooled. Figure 24 illustrates a Western blot of the stock preparations of separated RV El, E2 and Cd, each antigen preparation is pure in respect to viral protein contamination.  EC  11.0  E  C U,  U C  a 0  E  C  10  20  40  SO  160  320  R.ciprocal DIutIon of MtIgen Preparation  Figure 25. Comparison of Optimal Coating Concentration for El, E2 and Cd. Anti-RV protein IgG EUSAs were performed with a constant concentration of anti-RV reference serum on microtitre plates coated with serial dilutions ofEl, E2 or Cd containing 0.1% SDS (•ff•) and followingremoval of SDS by KCI precipitation (I(From: Mauracher et al., 1992a) -70- of Protein Binding Polystyrene, in the format of 96-well microtitre plates, was chosen as the solid phase for the adsorption of the purified RV proteins. Different plates supplied by Nunc, Costar, Falcon and Dynatech were tested for their abilities to bind RV proteins under various coating conditions: PBS, phosphate buffer pH 7.4, carbonate buffered saline pH 9.4, carbonate buffer pH 9.4, distilled water and drying protein onto solid phase in distilled water. Immulon-2 plates in combination with PBS as the coating buffer were found to be optimal for the coating of each RV protein. Each protein preparation had an optimal coating concentration (Fig. 25) and the dilution yielding peak absorbances were established to be: 1:40, 1:80 and 1:20 for El, E2 and Cd respectively. The titration of the El and E2 preparation exhibited decreased absorbance values at concentrations of SDS of 0.01% and 0.005% ,whereas Cd was not affected by the presence of SDS. When SDS was removed from the antigen preparations by KC1 precipitation prior to coating, the marked prozone seen at dilutions of 1: lOand 1:20 was not observed. However, a loss of up to 50% of the protein was realized during KC1 precipitation. The decrease of El and E2 antigen signals at low dilutions may be due to the competition of SDS with protein for hydrophobic binding sites on the plate or actual electrostatic repulsion of SDS-protein complexes from the charged polystyrene surface. of anti-El, -E2 and  -c IgG by ELISA  IgG directed to the individual RV proteins was determined in a similar manner to that directed to whole RV. The laboratory standard serum containing 943 lU/mi of anti-RV IgG was analyzed on non-reducing blots and the ratios of IgG directed to each protein were determined.  The 943 lU/mi of anti-RV activity were then  assigned to anti-El, -E2 and -C IgGs according to the percentage determined by Western blot. Figure 26 shows a Western blot of the standard serum with its densitometric trace. The dimeric forms of the envelope proteins, El dimer (El d) and El-E2 heterodimer make up about 35% of the total signal on the blot, and were included in the calculations. The slowest travelling dimer band was assumed to be Eld. A second dimer band, below the El d’ often stains more weakly, and was presumed to be the El -E2 heterodimer. The signal from this band is attributed to both anti-El and -E2 IgG and was divided according to the ratio of El/E2 monomer specific IgG. E2 homodimers has thus far not been described. It is not known if distinct epitopes exist for the dimeric El or E2 forms, but all the described El- and E2-specific mouse monoclonal antibodies show reactivities to both -71-  monomeric and dimeric forms of El or E2 (J.Wolinsky, University of Texas, personal communication). The standard serum was blotted and scanned in four separate experiments and the serum was assigned the following concentrations of RV protein specific antibodies, expressed in arbitrary units per millilitre of serum (AU/ml):  El (456 AU/mI), E2 (390 AU/nil) and C (97 AU/nil).  Five dilutions of the standard serum were  included on all plates coated with either detergent solubilized RV or electrophoretically separated El, E2 or Cd. The concentration of protein-specific IgG could then be determined by performing quadratic regression on the absorbance values obtained for the dilutions used for the construction the standard curve.  V  (i  1.4 1.4  1.0 0.8 0. 0.4 0.2  0.c  I  I  I  Eld El-E2  I  it  it  Cd  El  E2  Figure 26. Western Blot and Densitometric Scan of Anti-RV Standard Serum. 48.4% of total signal strength was attributed to anti-El IgG (13.9% from Eld, 9.6% from El-E2 and 24.9% from El); 41.4% due to anti-E2 IgG (11.5% from El-E2 and 29.9% from E2) and 10.2% due to anti-C IgG (10.2 from Cd). The El-E2 heterodimer can be seen as a doublet on the blot, but is poorly resolved, as a peak with a shoulder, on the desitometric scan.  3.2.3.Antipenicity and Bioloical Activity of Separated RV Structural Proteins As described above it was established that the antigenicity of the El and E2 proteins were sensitive to denaturation by reducing agents.  It was important to determine the effects of SDS and heat, the main  denaturants involved in non-reducing SDS-PAGE, on the antigenicity of the separated RV proteins. Antibody quantitation and affinity measurements were dependent on intact proteins, which had retained their epitopes during the purification protocol. Initial results indicated that denaturation by low concentrations of SDS (0.25%) was reversible but that boiling  -72-  of RV in non-reducing sample buffer lead to irreversible denaturation leading to a loss of antigenicity as measured by EUSA (Fig. 27). Whole RV preparations were diluted 1:10 in a 0.25% solution of SDS and then either boiled for 3 minutes or Left at RT. As a control, a preparation was diluted in PBS only. All preparations were then serially diluted from 1:10 to 1:1280. The control titration curve is non-linear at dilutions of 1:10 and 1: l6Oand becomes linear beyond that. The virus, treated with SDS without heating, showed an equal end-point titre to the control preparation, indicating that denaturation by SDS was reversible. In the saturated region of  n—c No SOS u—n .25 % SDS -c .25 % SDS, boiled  2.0  E  C  10  20  40  80  160  320  640  1280  Viral Antigen Dilution (1/x)  Figure 27. Effect of SDS and Temperature on the Antigenicity of Whole RV. RV preparation were either not treated with SDS (0) or treated with SDS (0.25%) and subsequently held at either 60#C (1) or boiled () for 3 minutes. Each RV treatment group was coated in serial dilution onto microtitre plates and its antigenicity was tested by RV IgG-specific ELISA, using a constant dilution of human anti-RV reference serum (From: Mauracher et al., 1992a).  the titration curve a noted depression of absorbance was observed indicating that SDS at concentrations above .001% interfered with binding of protein to the solid phase.  Heat denaturation in presence of SDS led to  irreversible loss of antigenicity indicated by a lower end-point titre of this preparation. This treatment reaches the same plateau as the unboiled SDS-treated preparation in the saturated region of the titration curve, again suggesting that the loss of absorbance is due to binding inhibition rather than denaturation. Although the above  -73-  experiment indicated that denaturation of RV proteins due to the employed SDS-PAGE conditions (0.25% SDS, 60C, non-reducing conditions) were minimal, the effect of separation on the biological activities of El and C protein were further investigated to determine if the purified RV proteins had retained their antigenic integrity. Biological activity of a protein is generally dependent on its structure.  It was therefore reasoned that if  separated viral proteins had maintained their biological functions, then their structure and also their structurally dependent epitopes had remained intact throughout the purification procedure.  Both El and C protein  preparations were analyzed for some of their biological activities. of El The El protein of RV contains the majority of the currently identified neutralization domains (Waxham and Wolinsky, 1985b) as well as the region involved in the agglutination of chick and goose erythrocytes. Both these activities are sensitive to denaturation (Ho-Terry et al., 1985; Qui and Gillam, 1992). This protein is not only thought to be involved in attachment of the virus to the target cell, but is also involved in low pH induced virus endosome membrane fusion, through low pH (<5.5) activation of fusogenic properties (Katow and Sugiura, 1988). Separated El protein preparations were analyzed for their HA- and HAl antibody binding activity. The separated El protein by itself was not capable of hemagglutinating chick red blood cells (CRBC) (Dr. S.Gillam, Dept. of Pathology, UBC, personal communication). This was not surprising as CRBC agglutination requires whole virion or cross-linked El.  The HA activity of the electrophoretically separated El was measured  indirectly by its capability to compete for HAl antibodies with native Rubella virions (Fig. 28).  The El  preparation was capable of competing for HAl antibodies in human serum. El preparations at a volume of 25 and 12.5 11, were capable of binding sufficient antibody in the 1:160 diluted antiserum to allow virus to crosslink CRBC and form a mat. At volumes of 6.25 11 or less of competing El, sufficient HAl IgG was left to inhibit CRBC crosslinking. The equivalent series using C preparations showed that the competition for HAl antibodies was exclusively due to El. As HAl antibodies recognize structural domains on El (Ho-Terry et al., 1985), it was possible to conclude that the separated El had maintained its native structure during purification procedures. An alternate approach for directly measuring intact HA activity of El preparations would have been to let the KC1 treated El preparation interact with CREC followed by the exposure of such sensitized erythrocytes -74-  Volume of lnhibftor 25  1 :40  E  1:80 El 1:160  >  1:320 0.  E 1:40 0 C  .2 1:80 0  C 1:160  12.5  6.25  3.1  25  0  12.5  6.25  3.1  0  0 00 00: 0000000 0000000 of 0000000 0000000 0* OG00000 0000)  Neg. HAl Serum, 1:8  Neg. HAl Serum, 1:8 Serum Control Cell Control  ooooo  1:320  Figure 28. HAT Antibody Competition Assay. El and C preparations were used to compete for HAT antibodies with whole rubella virions. Four dilutions of the anti-RV standard serum (1:40, 1:80, l:160,and 1:320) with a HAT titre of 1:160 were added to 5 wells per row. To each row, containing constant serum dilutions, varying amounts of either El or C preparations were added (25, 12.5,6.2,3.1 or 0 11). Each well was then brought to a fmal volume of 50 11. After a 1 hour incubation at 44C each well received 25 11 of 4 HA units of RV. Following 30 minutes incubation each well was loaded with 25 11 of a 0.02% solution of CRBC. Negative controls included the use of RV negative serum at l:40with both El and C antigen. Serum controls tested both negative and positive serum for endogenous HA activity. Cell control tested for button formation in the absence of both serum and HA antigen.  to HAl positive antisera in the presence of complement.  Lysis of the cells would have indicated binding of  Elonto the erythrocytes, thereby proving biologic activity of the HA domain of El. of C protein The biological activities of the RV C and E2 proteins are less well defmed. Although E2 protein was recently described to contain both N- and 0-liniced oligosaccharides (Lundstr6m et al., 1991) and is proposed to be an absolute requirement for the transport of El to the cell surface, followed by virion assembly (Hobman, 1989),  -75-  a clear biological function of E2 in the assembled virion has not been established. Hence, it was not possible to analyze the biological function of purified E2 preparation.  Similarly the capsid protein of RV has no  described biological function in the assembled virion, other than the described morphological feature of the nucleocapsid. By inference, from the work on the related aiphaviruses, SFV and SV, it is proposed that the C protein initially binds to the 5’ end of the full length + sense RNA (Weiss et a!., 1989) to initiate nucleocapsid assembly, leading to the budding of the assembled nucleocapsid from the plasma or trans-Golgi membrane (Hobman, 1989). During experimentation on the detergent extraction of RV envelope proteins, using TX-i 14, it was noted that the capsid protein, which under physiological pH conditions is not soluble in detergent, became detergent-soluble at pH 4.5 (Fig. 29A). This change in solubility properties was examined further using detergent extractions of  pH EDTA—  +  —  45  9.6  7.3  45  Triton X-114 Pellet  RuV  Sucrose Cushion  A  +  —  +  —  9.6  7.3 +  —  +  —  +  B  pH ,  ,,a.  mA  4  5  55  6  6.5  7  8  V  —  El  El  E2  E2  c  C  Figure 29. Autoradiographs of TX-i 14 Extracted RV. L-[ S]methionine-labelled RV in TCS was detergent 35 extracted and separated by reducing SDS-PAGE. A: Extraction performed at pH 4.5,7.3 and 9.6 ,in the presence and absence of 10 mM EDTA, showing the detergent pellets and the corresponding 6% sucrose cushion. The lane marked RuV contains untreated rubella virus. B: Detergent pellets from TX-i 14 extraction performed at p1-I 4-9 are shown in the first seven lanes from the left. Lane V contains untreated rubella virus. (From: Mauracher et al., 1991a) TCS containing radiolabelled RV performed at narrow pH intervals (Fig. 29B). C protein underwent a solubility shift from being hydrophilic to hydrophobic below pH 5.5. RV proteins are not acid precipitated at pH 4.5,nor does Triton X-114 change its properties within the pH range used in these experiments. The appearance of the -76-  capsid band in the detergent phase below pH 5.5 can be explained by a p11-induced structural change of this protein, either causing an intrinsic change in its solubility properties or allowing it to interact more strongly and co-solubilize with the lipid soluble El. To rule out the latter, TX-i 14 extractions were performed on cell lysates of insect cells expressing the cloned RV capsid gene, independent of El, E2 or the RV RNA (baculovirus expressed C protein was generously supplied by Dr. S.GiIlam, Dept. of Pathology, UBC).  This recombinant  capsid protein also underwent the solubility shift described for the wild-type capsid, indicating that the proposed solubility shift was an intrinsic feature of the C protein (Mauracher et a!., 1991a). To establish that the observed capsid solubility shift also represented a property of the intact nucleocapsid, the effect of TX-i 14 extraction and SDS treatment at pH 5 and 7 on the sedimentation of the virion and nucleocapsid through 22% sucrose was determined (Fig. 30A). The relative concentration of virion pelleting through the sucrose cushion following treatment at pH 5 or 7 in the absence of detergent was compared to the equivalent volume of TCS loaded directly onto the gel. The pronounced “smile” and lowered apparent MW is caused by the displacement of El by large amounts of BSA present in the TCS. TCS treated with 1% SDS at both neutral and acidic pH resulted in the disassembly of both the virion and the capsid into individual structural proteins, which remain buoyant in 22% sucrose, resulting in the absence of viral proteins in the pellet. Intact capsid can be differentiated from disassembled capsid by its ability to pellet through the 22% sucrose cushion. Extraction of TCS with 1% TX- 114 at neutral pH did not interfere with the capsid structure as C protein was detected in the pellet of the ultracentrifuged aqueous supernatant of the pH 7 extraction. If the TCS was acidified following the addition of TX-i 14, capsid was removed by its solubility in the detergent phase, resulting in the absence of a detectable C protein in the pellet upon ultracentrifugation. Viral RNA was detected in the pellets of all treatment groups (Fig 30B). The naked 40S genome of RV pellets through 22% sucrose (Oker-Blom et al., 1984). The presence of RV genome in the pellet following TX-i 14 extraction at low pH indicates that RNA had been released into the aqueous supernatant following the solubilization of the proteinaceous structure of the nucleocapsid into the detergent phase. Recent observations made on the cytopathic vacuoles of SFV infected cells (Froshauer et a!., 1988) have shown that the virus-encoded RNA polymerase remains associated with the limiting membrane of these vacuoles and that ribosome-RNA-capsid complexes form upon fusion of the viral envelope. This would be consistent with a proposed acid-induced intra-endosomal uncoating strategy, where the solubility shift of the capsid protein at -77-  A  SDS  MW 7  5  TX-tM  None  7  7  5  TCS Vp  5  pH  .-  —  84DEl E2  58D-  C  26  B  M  None 5  TX-114  75  SOS  Neg.  757  Figure 30. Ultracentrifugation of RV Following Detergent Extraction A: Western blot analysis of resuspended pellets following ultracentrifugation. TCS containing i0 pfulml of RV was treated with I % SDS at pH 5 or 7 p1-I 5 or 7) or with no detergent at both pH 5 or 7 (None, pH 5 or 7). overlaid onto 22% sucrose and centrifuged at 39’OOO rpm. Alternatively TCS was brought to a final concentration of 1% TX-i 14 and then acidified or left at neutral pH. Following extraction the aqueous supernatant was overlaid onto sucrose cushion and centrifuged (TX-i 14, pH 5 or 7). The lane labelled Vp contains 0.1 11 purified RV, lane TCS contains the equivalent amount of RV that was loaded onto the gel for each of the above described treatment groups. Molecular weights (x 1000) are indicated on the left margin, and the identity of the RV proteins on the right. B: Visualization of RV-specific PCR products. cDNA synthesized from the resuspended pellets of the six treatment groups described above was amplified by PCR using El-specific primers delineating a 287-hp fragment. Lane labelled M contains HaeIJI digest of pUC 19. Lane labelled Neg. contains a negative control for contaminating DNA sequences. (From: Mauracher et al., 1991a) ,  ,  pH 5.5 leads to the disassembly of the capsid within the endosome and the release of the viral RNA (Fig. 31). This concomitant shift in structure and properties of both El and C below pH 5.5 suggests that the solubility shift of the capsid plays a role in capsid uncoating.  It has been questioned whether the acidification of the  endosome may necessarily lead to the exposure of the nucleocapsid to an acidic environment (Kielian and + Helenius, 1986). These authors reasoned that the viral membrane remains impermeable to H ions and that  the endosomal  milieu would, therefore, not affect nucLeocapsid integrity.  Experiments using ion flux  measurements across artificial bilayers upon SFV membrane fusion have shown that ion permeability of the viral membrane occurs only after injurious events such as sonication, freeze-thawing or storage of the virus at RT (Young et al., 1983). These observations support the hypothesis that the intact viral membrane of freshly  -78-  Figure 31. Proposed Model for the Entry of RV into the Host Cell. Following attachment to the cell surface receptor (1) the virus is taken up via the receptor and internalized into a coated vesicle (2). The low pH in the endosome induces the El protein to become fusogenic (Katow and Sugiura, 1988) and facilitates the fusion between virus and lysosome membranes. In parallel, the low pH induces a solubility shift of the capsid, allowing it to associate with membrane, causing intra-endosomal uncoating (3). Upon membrane fusion, viral RNA is released into the cytoplasm. (From: Mauracher et al., I 991a)  harvested virions remains impermeable to small ions. However this group concludes that endocytosis in itself can compromise membrane integrity. Recent evidence suggests that the spike proteins in SFV act as an ion channel upon activation by the low pH environment of the endosome (Kempf et al., 1988). The current hypothesis on the entry of togaviridae suggests that within the endosome a low p11-induced conformational change of the viral spike proteins takes place resulting in a) a proton influx into the virion, b) the low pHinduced solubility shift of the nucleocapsid leading to the uncoating of the viral RNA and c) the fusion of viral and endosomal membrane (Mauracher et al., 1991a; Schlegel and Kempf, 1992). The investigation of the solubility properties has led to the formulation of a novel biological property for the capsid and has lead to a more complete understanding of the events occurring during early events in viral replication. The property of togavirus capsid proteins to undergo a structural shift leading to solubility changes and the unique b-barrel structure common to many RNA virus capsid proteins (Fuller and Argos, 1987) makes this event in viral replication a good target for drug intervention. Rhinovirus (picornaviridae) employs a similar -79-  entry strategy to the togaviridae, in that the low pH of the endosome leads to the uncoating of the nucleocapsid (McKinlay et a!., 1986). This event can be specifically inhibited by WIN 51711,a drug which tightly associates with the capsid and causes the protein shell to become less flexible, thereby preventing uncoating of RNA (Badger et al., 1988). It would be of interest to screen similar compounds for their ability to interfere with the solubility shift of the RV C protein in an effort to identify a chemotherapeutic route to interfere with the replication of RV and other togaviruses.  MW  pH5  pH7  TXAq  TXAq  58—  El •E2  3726—  C  Figure 32. Detergent Extraction of Separated C protein. Preparations of separated C protein were treated with 50 mM KC1 to remove SDS and then adjusted to pH 7 and 5 with 1 N HCI. Preparations were then extracted in 1% TX-i 14, and both the detergent phase (TX) and the aqueous phase (Aq) separated on 10% SDS-PAGE under reducing conditions. Lane RV contains whole RV, all other lanes contain pre-stained MW markers, indicated on the left margin.  The description of a function of C protein, that is structurally dependent, allowed the analysis of the electroeluted C protein. The C preparation was extracted in TX-i 14 in both acidic and neutral pH, following the removal of SDS by KC1 precipitation (Fig. 32). C protein retained its characteristic biological function and  -80-  became detergent soluble under acidic pH, suggesting that its structure was not irreversibly altered during nonreducing SDS-PAGE.  3.2.4.Summary Both Western blotting under non-reducing conditions and RV protein ELISA can be used to study the response of all immunoglobulin classes to the viral proteins. ELISA has the advantage of being faster, requiring fewer reagents and also providing easier interpretation of quantitative antibody measurement.  For the quantitation  and the assignment of units of protein-specific antibodies, care has to be taken to use both serum samples and the standard serum at dilutions in the linear region of the titration curve. Software programs which are capable of performing quadratic regressions and can transform absorbance data are of great value (ie: BioRad’s Microplate Manager  ,  especially if they can analyze data directly from the microplate reader.  Much of the work in the development of protein-specific ELISA was concerned with the determination of the optimal conditions for binding proteins to polystyrene microtitre plates. importance in maximizing signal to background ratios  ,  Although coating buffers were of  the crucial variable was found to be the type of  microtitre plate used in the ELISA. The presence of non-ionic detergent (ie: Tween-20), at concentrations lower than 0.05% ,in the sample buffer was also critical in obtaining good signals, particularly when using El and E2 protein preparations. It is possible that the non-ionic detergent replaces the protein bound SDS and allows the re-naturation of the protein on the plate surface. Of interest was the development of what initially appeared to be an insignificant observation, namely the pH induced solubility shift of RV C protein. One of the values of the academic environment is that it allows an individual to pursue an interesting phenomena on an intuitive basis. In the case of C protein the work lead to an additional study which culminated in the proposal of a new uncoating mechanism for RNA virus. Ultimately these extra studies became an important part in this thesis, by establishing that RV C maintained at least one of its biochemical properties following purification  -81-  3.3. Section ifi: RV-Specific Humoral Immune Responses in Adults Following Rubella  With the establishment of RV-specific antibody techniques it became possible to perform comparative analysis of the immune responses of patient populations following infection with RV.  The first subject population  studied was composed of adults who had a confirmed clinical and serological diagnosis of Rubella.  These  subjects will frequently be referred to as the “normal” or “control” study group. Subjects were randomly chosen from the study sample inventory list and were not selected on the basis of their clinical histories or the outcome of RV infection. It is necessary to define the normal RV-specific antibody response in an adult population in order to be able to compare and determine differences in groups developing pathological conditions followingRV infection, such as CRS or RAA. If such group differences in the immune response are detected, it may become possible to determine correlations with the pathogenesis or the etiology of the disorder. It is also important to establish the range in normal immune responses to RV as defmed in adults using current, as well as newly developed, laboratory techniques.  3.3.1.Response to Whole Virus and Separated RV Proteins The levels of IgG in a group of 15 adults who had confirmed clinical rubella at least six months prior to obtaining the serum sample, were determined using EUSA employing whole RV or individual RV proteins. The average value of IgG directed to whole RV was 202 JJJ/ml  —  65 (X  —  SD), which was in accordance with  previously reported anti-RV IgG levels (Dimechi et al., 1992). Currently RV-specific IgG levels A15 ITJ/ml are considered protective (WHO, 1971). Hence all individuals tested in this group are considered immune to RV re-infection. The response to El was dominant (208 AU/mi measured responses to E2 (56 AU/mi  —  .-  16) or C (33 AU/mi  115) and significantly (p=0.005) higher than the —  18) (Fig. 33).  3.3.2.Bioloical Activity of Sera Patient sera shown to be RV positive by the above described EUSAs were also analyzed by HAl assay. The median titre for this population was 1:32, with values of <1:8 being interpreted as negative (U. S. Dept. of  -82-  Health, Education and Welfare, 1970). The comparison of serological data obtained by traditional HAl  Control Rubella PaUents a  500  E  400.  300-  E  -r-  D -2000  I +  H•  T + I’  _.1__  _100.  I  0  Whole RV  El  E2  C  Figure 33: Quantitation of IgG to Whole RV and RV Proteins. Levels of IgG measured by whole RV EUSA are expressed in lU/mI, whereas levels of IgG to separated RV proteins are expressed in AU/mi. The 1SD. Values for individual patients are indicated by horizontal bars indicate the mean, error bars indicate dots. -  techniques and by solid phase irnmunoassay were in agreement.  Similar experiments, performed on a larger  scale (Herrmann, 1985) showed a good correlation between HAl and ELISA titres, leading ultimately to the substitution of HAT by ETA. Although these two approaches for the detection of RV-specific antibodies give similar results in this patient population, the data collected during this study show that the presence of high levels of anti-RV or anti-El immunoglobulins does not necessarily correlate with the presence of biologically active antibodies detected by HAl (see Section IV).  3.3.3.Reactivity to Linear and Topographic Epitopes As described above, El and E2 were sensitive to denaturation by reducing agents. By performing Western blots it was possible to determine the distribution of reactivities to continuous and discontinuous (topographic) -83-  epitopes among anti-El, -E2 and -C IgG (Table 7).  El  l0%±4%  E2  46%±41%  C  125% ± 83%  Table 7: Percentage of Signal Remaining on Western Blot under Reducing Conditions. 14 convalescent patient sera were sampled on reducing and non-reducing blot, and analyzed by densitometry. The ODmm for each protein was detern-thied and ratios of protein specific signal strength determined for reducing/non-reducing conditions. The mean ratio one standard deviation for these 14 patients is shown. —  Conclusions from the data shown in Table 7 were that in normal adults responding to RV infection, the El specific response was mainly directed to topographic epitopes with only 10% of the antibodies recognizing linear epitopes. A similar predominance of El specific antibodies to topographic epitopes has been reported in mice injected with whole RV (Wolinsky et al., 1991). Antibodies directed to E2 were equally distributed between specificities for continuous and discontinuous epitopes, however individual variability was observed. Antibodies directed to the C protein appeared to recognize both reduced and non-reduced protein.  In many cases  immunoblot bands were stronger under reducing-conditions, suggesting that linear epitopes may remain inaccessible to antibodies when capsid is in its dimeric form.  3.3.4.Kinetics of the IG Response Using the Western blotting techniques, the kinetics of IgG responses directed to RV structural proteins was studied over the first years followingRV infection (Fig. 34). The response to El was observed to be rapid and could be detected within the first several weeks following the clinical diagnosis of Rubella. The response to this protein reached maximal levels in the months following infection and remains relatively constant in the years following infection, with a slight drop-off observed in the second year following infection. The responses to Cd were observed to be weaker and transient, with levels being maximal at three months. Responses to the E2 protein differed from El- and C-specific responses, in that they exhibited a slow onset, with weak responses first observed at three months- and plateauing occurring by 12 months post Rubella.  -84-  Why the response to the RV E2 protein exhibits a slow onset remains speculative. The virus is cleared from the blood stream within days of the onset of the rubelliform rash and is not detectable in biological fluids  E  *  D  0 C V  C  IL  o u 0.6-  E a  A  El E2 C  *  0  0.4-  C V L  C V  0.2-  0  0.0 0 Days  10 Days  3 Months  1 Year  2 Years  Time Post infection Figure 34. The Kinetics of the IgG Response to RV Structural Proteins. 9 subjects were followed throughout the first two years following Rubella. Serum samples at consecutive time intervals were analyzed by nonreducing Western blot. Each time point represents the average density of the protein-specific bands for 9 SD. The top figure shows the mean value for each protein at the given time interval, with a fitted patients curve drawn through each point. The bottom figure shows the SD for each point, representing measurements on nine patients. —  -85-  shortly after the onset of the IgG response; it therefore is difficult to account for a continued B-cell response to E2, long after the virus has been cleared. It is possible that this viral protein may be retained in lymph nodes, as described for other protein antigens (Szakal et al., 1989), or that replicating virions persist. This however does not explain why only the E2 response is specifically stimulated in the months following infection. More intriguing is the hypothesis that continued stimulation of E2 clones is driven by cross-reactivity to self determinants.  Reactivities of El- and E2-specific T cells have been reported to be involved in autoimmune  disease in animal models (Yoon et al., 1991) and amino acid sequence homologies have been reported for E2 and the human proteolipid protein of myelin (Wolinsky, 1990). Direct evidence for crossreactivity of E2 specific antibodies to self constituents has remained elusive. Recent evidence has demonstrated that RV specific antisera crossreact with components of human synovial cells (Lund and Chantler, 1991). Therefore it would be of interest to investigate if the slow increase of anti-E2 IgG is associated with synovial tissue cross- reactivity. The analysis of clonal distribution of E2-specific B cells may support this hypothesis, as it would be expected that fewer clones would be detected in the late response if self-constituents were indeed responsible for the ongoing E2-specific antibody response.  3.3.5.Relative Affinity of IG directed to RV and RV Proteins Sequential serum samples from seven individuals undergoing serologically confirmed primary RV infection were collected at time periods of 10-20 days, 60-90 days, 360 days and 720 days following the onset of a rubelliform rash. Relative affinities of these sera were determined at each time point for IgG directed to whole RV and the individual RV structural proteins. Mean IgG ER(%) values  (— 1 SD) are shown in Figure 35. Measurement  of IgG anti-El affmity in acute phase sera showed initial low ER( %) values of 25% development of intermediate affmity values of 52% affinity levels of 85%  (-‘  (— 4%) with subsequent  11 %) at 3 months post-rash and final maturation to high  (— 7%) and 87% (— 3%) at 1 and 2 year intervals, respectively. In contrast, IgG anti-E2  and anti-C responses exhibited minimal increases over the two year post-infection follow-up period. In contrast relative affmity of anti-E2 and anti-C did not increase significantly (p “0.3) over the two year period investigated, with ER( %) values measured at 20 E2 and 21  (— 8%) in the acute phase and 31% (— 8%) after 2 years post infection for  (— 10%) to 36% (-‘8%) for C respectively. Measurements of IgG affinity to whole RV corresponded  to the results obtained with purified El protein, with ER(%) values increasing from 23% -86-  (— 4%) in the initial  10 days post-infection, to 52%  (— 11%) at 3 months, to significantly (p  and 87% (—3%) at 1 and 2 years following Rubella.  \0.000l) elevated levels of 85%  (— 7%)  Striking differences were noted in the maturation of IgG  affinity responses to individual RV structural proteins.  Whereas IgG affinity to the El envelope protein  paralleled results obtained with whole RV, IgG responses to the E2 and C protein underwent minimal affinity maturation, remaining at low levels throughout the 2 year post-infection period. The affinity response detected in the anti-El lgG is supportive evidence that this protein is important in eliciting both neutralizing antibody responses (Waxham and Wolinsky, 1985a) and may represent the main target for the cell-mediated immunity also (Chaye et al., 1992). The basis for the immunodominance of El remains unclear but may reflect the ability of antigen-processing cells to present the individual viral proteins to the human immune system (Johansson et  1  —S  0 -4-,  0 C 0 -4-,  D U  10—20  60—90  360  720  Days Post Onset of Rubefla Rash  Figure 35. Differential IgG Affinity Maturation to the Structural Proteins of RV. Elution ratios of IgG directed to each of the viral antigens El, E2 and C, as well as whole RV, were determined at sequential time intervals following diagnosis of clinical rubella in seven subjects. Each colunm represents the average ER(%) determination 1 SD for the seven patients. (From: Mauracher et al., 1992a) —  -87-  al., 1987). The high level of glycosylation of E2 with N- and 0-linked sugars, both containing sialic acid (Lundstr 6m et al., 1991), may be associated with increased resistance to proteolytic cleavage and subsequent reduced binding to the MHC class II complex (Sharon and Lis, 1982). In addition, both E2 and C may be less physically accessible within the intact virion. Capsid is an internal viral protein, while E2 may be masked by carbohydrate shielding (Klenk, 1990) and by virtue of its physical relationship with the El protein in the formation of surface spikes (Ho-Terry and Cohen, 1984). As it has been reported that B-cells specific for internal or shielded viral proteins receive less effective help from virus-specific Th cells (Scherle and Gerhard, 1989), it is possible that a lowered affmity of the E2 and C protein is due to decreased antigen-specific stimulation and Th support. Alternatively, the low affinity response to E2 and C could be explained by protein denaturation and subsequent loss of antigenicity during purification procedures or during binding to the poiyvinyl solid phase (Friguet et al., 1984). This possibility seems unlikely as it would be expected that El, being most susceptible to denaturants, would most easily be affected by alterations of high affmity binding sites. Furthermore it was shown in a previous section that by a variety of techniques denaturation of the separated El and C proteins was minimal.  3.3.6Summary and Conclusions The findings on the humoral responses to RV in adults are summarized below: L El-Specific Response: the antibody response to this protein is immunodominant and significantly (p = 0.005)  higher than antibody responses to E2 or C. The kinetics of the anti-El response is characterized by a rapid onset in the first several weeks following infection and is maintained in the years following Rubella. The affmity of these antibodies is high. Anti-El IgG reacts mainly to topographic epitopes (90%), with only 10% being reactive to linear epitopes. Both IgA and 1gM are reactive to this protein. L E2-Specific Response:  the antibody response to E2 is significantly lower than the El-specific responses.  Kinetic studies showed that the onset of these antibodies is slow, reaching maximal levels in the first year following Rubella.  The IgG affmity remains low, anti-E2 IgG is reactive to both topographic and linear  epitopes. L C-Specific Response: IgG responses against this protein are weakest and are transient. Peak reactivity occurs  -88-  in the months following Rubella and falls of in the years following infection. Similar to E2-specific responses the affinity of IgG directed to C remains low. Epitopes on this protein are not sensitive to denaturation and it appears that the majority of anti-C IgG is reactive to linear epitopes. Responses to RV Proteins The Western blotting technique had as its primary advantage an increased specificity over immunoprecipitation and the ability to analyze the IgG-,IgA- and 1gM-specific responses to RV structural proteins (see Fig. 22). In the investigation of the kinetics of the IgA responses of 33 male and 67 female adults over the years following RV infection it was observed that only females produced IgA antibodies specific to RV E2 (Fig. 36) (Mitchell et al., 1992). No significant differences between male and female were detected in the protein-specific responses of either IgG or 1gM. The clinical implications of differences in the recognition of E2 by serum IgA between males and females remains unclear. It was speculated above that E2 may have possible cross-reactive epitopes with self-determinants and this taken together with the well-known increased incidence of RAA in females may suggests an involvement of the serological responses of E2-specific IgA in the initiation of joint inflammation following Rubella or RV immunization. 1-4 Weeks  0-6 Days  El  E2  C  C  1-3 Years  1-11 Months ,•//  7/ IC IC ‘C  IC CO to  Ii — El  E2  .. C  1011 t:_  El  E2  C  Figure 36. Kinetics of the Appearance of RV Protein-Specific IgA as Determined by Western Blot. The kinetics of appearance of El-, E2- or C-specific IgA were determined in males (checkers) and females (hatched) patients. Sera were collected sequentially from each patient during the early acute (0-6 days post onset of rash), acute (1-4 weeks), early convalescent (1-11 months) and late convalescent (1-3 years) phase of rubella infection. Specificities were determined by class-specific immunoblot assays. Results shown are the relative percentages of female or male patients at each interval. (From: Mitchell et al., 1992a) -89- Protein Specific LymphoproliferativeResponses  Following the characterization of antibody responses to the individual proteins of RV it became possible to compare these to lymphoproliferative responses.  The proteins obtained by preparative SDS-PAGE were of  insufficient concentration when used by others in stimulation assays with human lymphocytes (Dr. D.K.Ford, Arthritis Society, Vancouver; Dr.D.Ou, Dept. of Pathology,UBC). A collaborative study was therefore initiated (Ms.H.H.Chaye and Dr.S.Gillam, Dept. of Pathology, UBC) to compare both B and T cell responses in 14 randomly chosen adults, without previous knowledge of their immune status to RV.  IgG responses were  analyzed by protein-specific EUSA and T cell responses were determined by using RV El, E2, and C obtained from a vaccinia virus expression system in lymphoproliferative assays (Chaye et al., 1992). In the latter assay stimulation indices (SI) greater than three were considered sigiiflcant. The comparison of the B- and T cell responses in these patients is illustrated in Figure 37.  Serological Response  Proliferative Response  •  E  :  4<  2  •  C  I  -  0  1.  -  1  29  I  I  El  E2  El  C  0  I  I  --  E2  C  Figure 37: Comparison of Proliferative and IgG Responses to RV Structural Proteins. Values of stimulation indices (SI) for the proliferative responses and A.U./ml for the protein specific IgG responses are on a natural the standard error. Proteins used in the serological study were obtained by log scale. Error bars indicate preperative SDS-PAGE from whole virus, whereas proteins used in the lymphoproliferative assays was derived from cloning RV protein genes into vaccinia virus expression vectors. (From: Chaye et al., submitted) -  -90-  The comparison of the RV-specific serological and lymphoproliferative responses in this population showed similarities in their protein specificities. The response to El was immunodominant in both responses over the reactivity to E2- and C-specific responses, again suggesting that the response to the El protein plays a paramount role in the induction of protective immunity. These data suggest that the El protein may be the best candidate for developing subunit vaccines for RV, as it gives rise to both a vigorous antibody and lymphoproliferative response. Investigation of the precise regions of El that are involved with neutralization and hemagglutination have been reported to lie between an residues 244 and 300 of the El protein (Terry et al., 1988). Synthetic peptides constructed from this and other region’s primary sequence are currently being investigated as potential B- and T cell recognition sites and may lead to the assignment of dominant T-cell epitopes as well as the sites of continuous B-cell epitopes on El. This approach may ultimately identify small regions from the El protein which can be used to induce a protective immune response to RV, thereby avoiding RV epitopes which may be involved with adverse vaccine side effects. Observations concerning the sensitivity of the El protein to denaturation by both detergents and reducing agents have revealed that, by far, the majority of the El and E2 epitopes are topographic and structurallydependent.  The use of native antigen will therefore be an important consideration in the development of  sensitive immunoassays using bio-engineered RV proteins. The development of expression systems allowing the large scale production of pure, correctly folded and glycosylated RV El is currently under investigation at many centres.  The use of baculovirus as an expression vector for the production of cloned RV El and E2 in  Spodoptera frugiperda cells has been reported (Oker-Blom et al., 1989; Seto et al., 1991) and is proposed to provide a more economical alternative to the use of whole RV virus in immunodiagnostic assays (Sepp dñen et a!., 1991).  -91-  3.4. Section IV: RV-Specific Humoral Immune Responses in CRS Patients  Congenital Rubella Syndrome (CRS) patients and their immune status have been investigated since the initial description of this syndrome in 1941 (Gregg, 1941). Although this manifestation of RV infection has been dramatically reduced in the developed world by the implementation of a successful vaccination programme, and has therefore lost much of its clinical urgency, the disease remains of interest as a model for viral persistence and tolerance induction.  The infection of the human fetus with RV early in gestation, clinically presenting as  CRS, leads to long term persistence of the virus in these individuals. Virus persistence is thought to result from RV-specific tolerance although little direct evidence has been presented to support such a mechanism for continued viral replication. It has remained a paradox why CRS patients mount a humoral and cellular response yet remain incapable of clearing the virus. To explain viral persistence, the affinity of anti-RV IgG (Fitzgerald et a!., 1988), HAl activity (Cooper et a!., 1971) or humoral and cellular responses to the individual RV structural proteins have been investigated (Katow and Sugiura, 1985; de Mazancourt et al., 1986; Chaye et al., 1992), in an effort to defme a partial tolerant state to RV arising from either selective anergy of neutralizing IgG, clonal deletion of T-effector cells, lowered concentrations of El specific antibodies or the inability to produce a high affmity antibody response. RV-specific antibody responses of CRS patients were compared to those of normal adult control patients, described in Section ifi. The observed differences in how CRS and Control patients responded to RV support the hypothesis that intrauterine RV exposure has led to virus-specific tolerance in CRS patients.  3.4.1.RV and RV-protein Specific ELISA CRS patients (n= 15) and adult Rubella control patients (n= 15) were compared for their quantitative IgG responses to whole RV and the separated RV structural proteins (Fig. 38). Whole RV-specific IgG antibody levels as measured by whole RV ELISA were 202-65 lU/mi (X -SD)for the control group, whereas the CRS group demonstrated lower levels at 124—55 lU/mi. However, these differences were not statistically significant at the p \0.05 level. No differences in the responses to the E2 and C protein were determined between the CRS and control group but significant differences were observed in the El-specific responses. In the control group,  -92-  the majority of RV-specific IgG antibody was observed in the anti-El response (208-115 AU/mi) which was In the CRS group, the anti-El IgG levels were  greater than either the E2- or C-specific IgG response.  considerably lower (47 -25 AU/mi) and significantly lower (p =0.004) than those measured in the control group.  Control Rubella Patients  CRS Patients  500  500  300  —  D ‘  300  T D  +  200  ‘‘  200  01  100.  I •  •  I  j  E2  C  0  Whole RV  El  100  Ii  0 Whole RV  El  E2  C  Figure 38: IgG Responses to RV and RV Proteins in Control and CRS Patients. Results are expressed in 1 Standard lU/mi for whole RV-specific IgG and AU/mi for RV protein specific IgG. Error bars indicate deviation around the mean. Values for individual patients are also shown (L). (From Mauracher et al., 1992c) -.  The RV-specific immune response of CRS patients in this group would normally be considered to be protective, as it was higher than 15 lU/mi of anti-RV IgG, the suggested cutoff value for the definition of protective immunity (WHO, 1971). Yet RV is known to persist in CRS patient populations and RV reinfections have been reported in school aged CRS patients (Hardy et al., 1970), which has been taken to be indicative of an underlying defect in the RV-specific response in these patients. The selective loss of El antibodies has been previously described  in a small CRS population using immunoprecipitation techniques (Katow and Sugiura,  1985). A subsequent study by others (de Mazancourt et al., 1986) did not report such a decrease in El specific -93-  IgG. However, as the CRS patients studied were considerably younger (\1 year), maternal IgG may have interfered with determination of autologous RV-specific IgG. Furthermore, no steps were taken to correct for the unequal percent distribution of methionine in the RV structural proteins when 35 S methionine labelled preparations were used in their irnmunoprecipitations. The above study confirms Katow’s and Sugiura’s fmdings and suggests that CRS patients have significantly reduced levels of El specific antibodies.  Although anti-El  levels were reduced, the majority of CRS patient did have circulating anti-El IgG. Therefore the biological activity and affinity of these antibodies was investigated further.  3.4.2.Biological Activity of Sera in CRS Patients HAT antibody levels were determined in sera from 11 CRS patients ranging in age from 1 to 27 years of age and were compared to those of control patients described in section ifi. The median HAT titre of the control group was 1:64 and compared to a median titre of (1:8 in the CRS group.  HAT values in the CRS patient  group did not exhibit a normal distribution and the two populations were therefore compared using nonparametric statistical analysis (Mann-Whitney U test) for the 1/log . The HAl titre showed significantly lower 2 titres in the CRS group as compared to the control group (p=O.003). A lack of biological activity of the anti-RV IgG response in CRS patients, as measured by HAl, has been previously described (Cooper et al., 1971; Ueda et al., 1975). Whether this correlated with a concomitant lack of in vitro neutralizing antibodies was not determined.  In the few CRS patient samples for which both NT and HAl tests were performed in the  laboratory, the two tests gave coinciding results, with negative HAl titres correlating to negative NT tests. However it has been demonstrated by the use of monoclonal antibodies, that HA and neutralizing domains do not necessarily represent the same epitopes (Brush, 1988; Waxham and Wolinsky, 1988). A possible mechanism for the lack of HAl activity of the El specific IgG in CRS patients is a lack of high affinity antibody. In many CRS patients it was observed that partial HAT was occurring, similar to the soft radial hemolysis pattern described for low affinity anti-RV IgG by the Radial Hemolysis typing test (Hedman et al., 1989). HAT titre determinations have been criticized because of inaccuracy in the two-fold dilution steps, and it has therefore often been suggested that any given HAT titre should be interpreted as a range. plus/minus one dilution from the determined HAT titre. Statistical analysis of HAT titres are nevertheless valid as the variance observed in  -94-  any given population reflects both random- and measurement error.  3.4.3.Relative Affinity of-RV Specific IpG in CRS Patients Affinity indices were determined for IgG directed to whole RV as well as to the separated structural proteins El, E2 and C in both CRS- and a control group patients (Fig. 39). Serum from adult control group (n= 5) showed high affinity IgG directed to whole virus and El, but had significantly lower affinity of IgG directed to E2 and C (p \O.OOl),correlating with the data presented earlier in Figure 35. The CRS patient population was subdivided according to age into a child (n=8, age range 1-4 years) and adult (n=7, age range 17-34) groups.  tuu  Whole RV 80  x 0  C  C  60  40 / / / /  20  0—-  Child Control  Adult  Child  CRS  Control  Adult CRS  x 0  C >‘  C 9-  Child Control  Adult  Adult  Child  CRS  Control  CRS  Figure 39. Affinity Indices of IgG directed to RV and RV Structural Proteins in CRS and Control Patients. 15 adult control patients in the convalescent phase of Rubella (white bars) were assayed on affinity ELISA. Results are expressed as the affinity index on the y-axis. A CRS population of 15 patients was divided into childhood group (hatched bars, n=8) and adult groups (cross-hatched, n=7). Error bars are SD. —  -95-  The latter group was age-matched to the control group. The affinity of IgG to whole RV was observed to be significantly lower in both CRS groups (p=0.00l) as determined by one-way ANOVA. The affinity of anti-El antibodies also was significantly reduced (p =0.0001) in both CRS groups if compared by ANOVA. The Fisher LSD test also showed significant differences in the means of both the child and adult IgG anti-El affmities when compared to the control group. No statistically significant differences were observed in the low affmity response to E2 in any study group (p =0.36). ANOVA analysis of the C-specific IgG affinity showed significant differences in the populations (p=0.0001). By using post-hoc analysis it was shown that the mean affmity of anti-C IgG in the adult CRS patients was significantly higher than the affinities of either the child CRS- or the control patient’s anti-C IgG. No differences were detected in the affinities of anti-C IgG of child CRS- and control patients. Antibody affinity is an important functional parameter of the humoral response and the inability to switch from low to high affmity antigen-specific responses has been an observed feature of immunological tolerance in animal models (Steward and Steensgaard, 1983). The inability of young CRS patients to produce high affinity IgG to El was marked.  Age may be a factor in the affinity maturation but it has been recently shown that  infants at age one can produce high affmity IgG (O.Meurman and K.Hedman, manuscript in preparation). The adult CRS patients of this study group appear to have partially regained the ability to produce higher affinity IgG to El and interestingly were observed to produce high affinity IgG to C protein. It is conceivable that continued antigen stimulation by persistent RV may have facilitated the maturation of the IgG responses in this patient group. The observed decrease in both quantity and affinity of the El-specific IgG in the CRS population may well be responsible for deficient antibody-mediated viral clearance mechanisms.  3.4.4.Reactivity to Linear and Topographic Epitopes in CRS Patients The presence of 2-mercaptoethanol reduces much of the antigenicity of the envelope proteins El and E2 but not that of C, by the destruction of epitopes dependent on tertiary protein structure (Mauracher et al., 1992a). In the analysis of CRS patient sera by reducing and non-reducing Western blots it was observed that these subjects were deficient in their ability to produce IgG reactive with reduced El. In contrast no group differences were observed in the reactivities of E2- and C-specific antibodies in these two populations (Table 8). This observation provided the most consistent feature which differentiated the CRS population from control responders. In respect to the other antibody parameters measured (HAl, affinity, anti-El levels) some CRS selective non-responsiveness to linear epitopes in CRS patients in CRS patients has been further investigated  -96-  Control Patients 4%  p-value  CRS Patients 0.5  \0.00l  1.1  El  10%  E2  46% —41%  31 —40  0.1  C  125% —83%  96 —54  0.5  —  Table 8. Percentage of Signal Remaining on Western Blot under Reducing Conditions for Control and CRS Patients. Control patients (n= 14) were compared to CRS patients (n= 24) for the fraction of their RV protein specific IgG reactive to linear epitopes.  patients were always found to lie within the range of control responders; in respect to antibodies to linearized El (El red)’ all CRS patients showed significantly reduced responses to linear epitopes in comparison to normal control patients. Immunoblot results of two representative patients are illustrated in the Figure 40. This  Normal Rubella  CRS  A Ai•  Non-Reducing  i 11  .,. 1• Cd El  E2  Cd El  -  E2  ‘I  Reducing  El  E2  E2  C  C  Figure 40. Western Blot under Reducing and Non-Reducing Conditions of Representative Control- and CRS Patients. Two blots, and their corresponding densitometric scan, are shown for a adult control patient and one CRS patient, Serum concentrations were equal for the reducing and non-reducing Western blots.  -97-  A Degradation by or DC  A  h  / El  E2  .  ()  Ig G  B-Cell  B-Cell  IgG directed to  IgG directed to  topographic El epitope  linear El epitope  Class II TCR  Clonal Anergy  Clonal Activation  Figure 41. Model for the Production of IgG to Topographic Epitopes in the Absence of El-Specific Th Cells. A hypothesis is proposed which explains the production the IgG response to topographic epitopes (A) and the lack of IgG responses to linear El epitopes (B) in the CRS population by a lack of El -specific Th cells. Th cells to El are proposed to have been tolerized in this population (crossed-out anti-El Th) due to exposure of the fetal immune system to RV. Scenario A depicts a particular clone reactive to a topographic epitope of El. The IgG product of this clone will be detected on whole RV/El protein ELISA or non-reducing Western blot. As the IgG recognizes the intact El, it will be capable of binding the intact virion encountered in the initial viremic phase. By phagocytosis and presentation of processed antigen in context of class II molecules, this clone will present peptides of all three structural proteins on its surface, and will in turn be able to receive the second activation signal by Th cells (Noelle and Snow, 1990) specific to the three viral proteins. The lack of anti-El Th cells does therefore not prevent clonal activation. On the other hand, the scenario illustrated in situation B, depicts a clone producing IgG to a linear epitope on El (vvv, partially or completely buried within the highly folded El protein. The most likely circumstance in which this IgG molecule will recognize its antigen is once the virus and its proteins have been degraded, as an outcome of local inflammatory responses or by exo-peptidases of dendritic cells in the germinal centres of the lymph node (Steinman, 1991, Liu et al., 1992). This clone will therefore bind partially degraded species of the El protein, and only present El peptides on its class II molecule. However, as Th cells to El are lacking, this clone will not receive any further stimulation and will remain in anergy. IgG directed to peptides or linearized El will therefore not be produced in patients lacking El specific Th cells. This model accommodates the serological findings in CRS patients and lends support to the hypothesis that interuterine exposure has led to immunological tolerance. (From Mauracher et al., 1992c)  -98-  by others using solid phase bound peptides representing the a sequence adjacent to the putative neutralizing and hemagglutination regions of RV El (Terry et al., 1988), between amino acid residues 213-239 of the RV El. It was demonstrated that Rubella patients as well as RV vaccinated individuals showed reactivity to this peptide in the weeks following virus exposure (Mitchell et al., l992b). The CR5 patients on the other hand did not show reactivity to this peptide which again suggests the inability of this population to produce IgG to linear El epitopes (Mauracher et al., 1992b, Mitchell et al., 1992b).  Is the non responsiveness of CR5 patients to  linearized El protein as well as El peptide based on an underlying deficiency of El specific Th cell responses and does this represent a specific tolerization of the El specific cellular response.  If so, the ability of CR5  patients to make IgG to conformationally dependent epitopes has to be explained. It is unlikely that these epitopes are T cell independent as El does not contain either multivalent epitopes nor are the carbohydrate moieties on this protein thought to be antigenic. A model to explain production of IgO directed to topographic epitopes of El in the absence of El-specific Th cells is proposed in figure 41.  3.4.5.Analysis of Sequential Serum Samples from CRS Patients Three CRS patients were retrospectively studied pre- and post MMR vaccination over a 2 to 6 year period. Results obtained following IgG in-miunoblot analysis of sequential sera from these patients are shown in Figure 42 and serological work-up is summarized in Table 9. The cord blood from patients A and B exhibited strong El responses in presence of both E2- and C-specific IgG as well as IgG specific to the reduced proteins; no cord blood was available for patient C. Cord blood contains maternal IgG at equal or higher titres than the maternal serum, and it was therefore not surprising that this sample contained reduced El reactive IgG. Following the first year of age, these patients lose the reduced El-specific IgG and HAl antibody responses which become undetectable in all three patients. Patients A and B did not respond to vaccine, with antibody levels remaining similar pre- and post-MMR vaccine. Patient C was unique in that no RV-specific antibodies were detected in the pre-MMR vaccine sample and the post-vaccine samples contained El-specific IgG only. However these antibodies were not reactive to linearized El nor did they show any HAl activity. Cord blood showed matemally derived HAl activity and high affinity El-specific responses which were then replaced by endogenously produced IgG of lowered affmity with no HAl activity.  -99-  A  Cord  i I  *  I  4,  ly  *  ly2m  ly6m  12m)  16m1  *  Age  3y 12yJ  post MMR  +  II  Cd El Elb. WE2 E2 -.  C  1: B  Cord  *  6m +  C  I  4,  2Y4m (lyl  +  +  1y7m 2y4m (8m1 —  +  *  ly5m (imi  —  +  3y8m  8y  12y1  (6y)  —  n  +  —  +  a  Figure 42. Western Blot Analysis under Reducing and Non-Reducing Conditions for Three MMR Vaccinated CRS Patients. Patients A, B and C were prospectively studied following MMR vaccine (arrow). A cord sample was available for patient A and B. Patient age is indicated on the top of each series, and time following MMR vaccination is indicated within square brackets [J. -100-  Patient B  Patient A Age  Birth  Post MMR:  ly 2m  ly Sm  2m  Sm  Birth  2y 4m  Patient C ly 7m  2y 4m 8m  ly  Who/c RV  77,9  51  48  77  60  /  16  El  81  19  29  78  44  /  32  E2  35  37  32  36  55  /  /  C  34  32  37  33  26  /  /  Whole RV  145  163  150  112  76  1  69  El  82  17  17  53  17  3  50  E2  24  58  41  25  50  3  3  C  17  19  20  10  5  1  1  HAl  1:32  1:8  <1:8  <1:8  SfltHElred  yes  no  /  no  Avidity index  igG (iUJmi)  1:8 no  1:32 yes  <  1:8 no  Table 9. Serological Examination of MMR Vaccinated CRS Patients. Patient samples in this table are indicated by an asterix on the Western blots in Figure 42.  3.4.6.Evidence for Th Tolerance in CRS Patients An immunological mechanism for the prolonged persistence of RV in the CRS infant remains speculative. The virus is not cleared from fetal or infant tissues although maternally transferred IgG, as well as in some cases autologous 1gM, can neutralize the virus in vitro. It is therefore likely that immunoglobulins are of little use in bringing a well established viral infection under control and that the cellular response is of primazy importance. The cellular response plays a key role in preventing primary viral infections in the human, as demonstrated by clinical frndings in X-linked agammaglobulinemia (XLA) and in congenital thymic aplasia (CTA) patients. XLA patients are deficient in IgG production but produce normal levels of T cells. These patients are usually diagnosed once their maternally derived IgG has been catabolized, late in the first year of life, resulting most often in the patient’s susceptibility to chronic bacterial infections. These patient also have difficulties in clearing  -101-  enteric viral infections (ie: polio virus) but have little difficulty in terminating systemic viral infections (Saulsbury et a!., 1980). On the other hand patients with complete CTA, incapable of producing T cell responses, often succumb to overwhelming viral or mycobacterial infections (Cooper and Butler, 1989). If cellular responses are of prime importance then the long term persistence of RV in fetal tissues can be readily explained, as maternal T cells do not cross the placenta and fetal T cells responses are largely suppressed in utero and for the first months after birth (Papadogiannakis et al., 1990). The gradual decline in the shedding of RV in CRS patients over the first two years of life may be due to the initiation of the endogenous cellular response to RV. Although CRS patients have been shown to have a specific deficiency in their lymphoproliferat ive responses to whole virus (Fuccillo et a!., 1973) others have reported that a non responsiveness to RV is strongly dependent on the time period of intrauterine exposure, with CRS patients who were exposed after 12 weeks of gestation showing near normal responses (Buimovici-Klein and Cooper, 1985). Non responsiveness of lymphocytes measured by whole RV proliferative assays remains unresolved, and the investigation of lymphoproliferative responses in CRS- and control patients using synthetic peptides representing RV proteins may help in a more definitive analysis of the T cell immunity. In the comparison of the B- and T-cel! responses of CRS and control patients to individual RV proteins, it was demonstrated that CRS patients showed strong lymphoproliferative responses to E2 and C, but were selectively deficient in their responses to El (Chaye et a!., 1992). As these experiments were performed with vaccinia virus expressed proteins, the background response to vaccinia in some of these patient may have caused false positive responses, in spite of steps taken to correct for vaccinia-specific signals. Preliminary results of lymphoproliferative assays using 30 an peptides representing the complete El, E2 and C regions have shown that 3 out of 4 CRS patients are completely non-responsive to El peptides, whereas all CRS patients react to E2 and C peptides. Of the control responders, 12 out of 12 adult control responders reacted to peptides from all three RV proteins (Dr. Dawei Ou, Dept. of Pathology, UBC, personal communication).  These observation support  the above hypothesis that the selective antibody non responsiveness to linearized El or El peptides represents an underlying immunological tolerance of the T cell response to El. Antigenic sites recognized by antibody may be regulated by Th cells and has been described in detail for the murine response to staphylococcal nuclease or sperm whale myoglobin (Berzofsky, 1985). Cognate interactions  -102-  between B- and T cells are thought to regulate T cell-dependent antibody responses, with the activated Th cell providing the secondary signal to the B cell, via class II-CD4 interactions (Noelle and Snow, 1990). The resting B cell can bind antigen by surface IgG which is thought to provide the first signal for cell activation. Antigen is then phagocytosed, degraded and re-expressed on the cell surface in context of class II molecules (Tony and Parker, 1985). Th cells, recognizing peptides originating from the phagocytosed antigen, can then bind to the B cell and provide the secondary stimulus to induce G 0 to G 1 transition of the B cell with fmal progression of the B cell to antibody secretion being promoted by IL-4 and IL-5. The existence of this pathway of B cell activation is supported by observations of a hierarchy of help for B cells responding to internal or external antigens of influenza virus in adoptive Th cell transfer experiments using athymic mice (Scherle and Gerhard, 1988). In this model B cells expressing antibody to surface antigens (HA or NA) internalize whole virion and can be activated by Th cells specific to either internal or external proteins (M, NP, HA or NA). In contrast, B cells producing IgG to internal proteins (ie: NP) can only be activated by NP-specific Th cells. It is believed that the B cell will encounter NP only when virions have been degraded, the NA-specific cell will therefore only bind and phagocytose NP antigen. A more recent experiment has described responses of mice which have been made transgenic for the G protein of Vesicular Stomatitis virus (VSV) (Zinkernagel et al., 1990). If these mice are immunized with cloned and expressed G protein no antibody responses are mounted. However if whole VSV virus is used as the inoculum IgG responses to all VSV proteins, including G protein, are made. This experiment was designed to further support the hypothesis that B cell tolerance is regulated at the Th cell level, but also demonstrates that B cells, producing antibody to viral surface structures, can be activated by Th cells with activities to other viral proteins. A similar scenario is proposed to exist in CRS patients where the individual, a primo to the stimulation of the RV-specific IgG response, has lost El-specific T cells as a result of intrauterine tolerization of the immune system. The model takes the observations made by Scherle and Gerhard one step further by differentiating not only between internal and external proteins of a virus but also between internal and external epitopes of the immune dominant viral protein (See Fig. 41).  -103-  .7.Summary and Conclusion 4 3. The responses of the humoral immune response in CR5 patients was found to differ from the response of adult control patients in the following manner: L No significant differences in the IgG responses between CR5 and control patients was detected when using  whole RV ELISA. L When these two populations were analyzed by RV protein-specific ELJSA, a significantly lower (p = 0.004) IgG  response to El was demonstrated.  No significant difference was shown in the E2- and C-specific response.  L The affinity of IgG directed to whole RV or El was significantly lower in the CR5 patients as compared to  adult control patients. L The CR5 patient population was characterized by their inability to produce IgO to linear El epitopes.  IgG  was produced to linear E2 and C epitopes. This is also reflected by the significantly reduced responses to El peptide (an 213-239), as demonstrated by others. L The CRS patients showed significantly lowered HAl titres. It appears that the El-specific antibodies present in this group are not biologically active, in the sense that lack HAl activity. L A comparative study of B and T cells responses revealed that both serological and lymphoproliferative  responses to El are significantly reduced in this population, as shown by Chaye et al..  The observations on the RV-speciflc humoral and cellular responses of CR5 patients identilS’ this population to be selectively tolerant to the El protein of RV. As this protein appears to be immunodominant and provides the main target for both serological and lymphoproliferative responses it is hypothesised that viral persistence in CRS patients is the outcome of the lack of a major part of the T cell response. Similar dominance of one protein acting as the main T cell antigen has also been reported in other viruses {ie: influenza A (Wraith, 1987) and BHV I (Hutchings et al., 1990)]. The reactivity of cytotoxic T cells in Rubella and CRS patients has not been investigated but it is hypothesised that protein specificities of the Tc cells are similar to that of the Th cells measured by lymphoproliferative assays (Thomson and Marker, 1989; Ou et al., submitted). Why the E2- and C- specific T cells may still be present in the tolerized patient population remains speculative, but the same properties which make these proteins poor T cell antigens may also make them poor toleragens. Cross-reactivity of T cell epitopes of RV capsid or the RV encoded nonstructural (ns) protein with other common viral capsid -104-  or ns proteins may also be possible. The capsid structure has been highly conserved in many RNA viruses and it may be possible that peptide sequences of RV C protein are shared with other capsid proteins or RNA binding proteins. However amino acid homologies between RV C protein and those of the related SFV and SV are between 10 and 20% only (Frey and Marr, 1988). These authors report that significant homologies exist in the nonstructural protein coding region for RV and that of many other RNA viruses. The ns proteins of RV have not been considered as T cell antigens but may potentially play a role in T cell immunity as is the case for DNA viruses (ie: EBNA in Eppstein-Barr virus). An amino acid sequence homology search for C protein of RV in GeneBank has not identified any sequence homologies with other viral capsid proteins but has identified short stretches of homology with bacterial and viral RNAIDNA binding proteins. The possibility of anti-RV B-cell activation by cross-reactive Th cells may therefore be valid. The fmding of predominantly low affmity anti-RV IgG in the CRS group may also be taken as indirect evidence for T cell tolerance. In the absence of a strong Th cell response it may be conceivable that B cells do not obtain sufficient stimulation to undergo multiple replication cycles during which affmity maturation can occur. This may explain not only low affmity antibody but also the observed fall in anti-El IgG titres in the CRS patients. Similarities exist in the description of tolerance in CRS patients and the perinatally LCMV infected mouse model. Murine LCMV membrane proteins contain conformationally dependent neutralizing domains (Wright et a!., 1989), yet the anti-LCMV IgG in these perinatally infected mice was not capable of clearing the viral infection. In this model, ineffective virus clearance correlated with low Tc cell responses and low delayed type hypersensitivity, but is unrelated to the levels of in vitroneutralizing antibody or NK cells (Thomson and Marker, 1989). Although IgG does play an auxiliary role in the rapid clearance of virus in healthy animals (Cerny et al., 1988), the persistence of LCMV in susceptible mice is thought to be solely due to absent LCMV-specific Tc cell (Oldstone, 1989). Nevertheless, virus specific Th cells must be present as IgG production to viral proteins requires Th support  -  LCMV infected athymic mice only produce 1gM specific to LCMV. Further it has been  well established that in the neonatally infected mice IgG directed to LCMV proteins are of low affinity and are commonly present as dC or as deposited as IC in the kidneys or other organs (Buchmeier et al., 1980). This response patterns shows striking similarities to the observations made on low affinity anti-RV IgG responses  -105-  (Fig. 39) and high levels of anti-RV IgG dC (Coyle et a!., 1982) in CRS patients. The protein specificity of the cellular response or IgG reactivity to linearized proteins have not been analyzed in LCMV tolerant mice and could reveal a similar scenario as in the CRS patients, where protective immunity may be dependent on the cellular response to a single protein. The selective non responsiveness of CRS patients to linearized El in their serologic responses may have relevance to the interpretation and future development of immunodiagnostic assays for Rubella and other systemic viral infections. CRS patients are characterized by a persisting RV infection, yet are seropositive with commercially available solid phase immunoassays and therefore are interpreted to have “protective immunity’. It is well established that neither the endogenously produced IgG nor the maternally derived IgG, having in vitro neutralizing activity, clears RV in the CRS infants. Therefore, it has to be questioned whether anti-RV IgG has a function in the clearance or the prevention of Rubella in vivo. It can be speculated that the only reason why the presence of IgG correlates with protective immunity in the healthy responders is because it reflects the presence of a functional T cell response. Serological assays for RV should therefore be designed to detect IgG which will correlate to the presence of Th cells specific to the El protein. As suggested by the model shown in Figure 41, responses to topographic and structurally dependent El epitopes may arise in the absence of El Th cells whereas IgG to linearized El are dependent on El specific Th cells. Therefore it might be more appropriate to use denatured El protein as the solid-phase antigen. Such an immunoassay would be capable of differentiating between the responses of control patients and the non-protective responses of CRS patients. Much effort is currently being invested to produce expressed El in native form, for example by the use of baculovirus expression systems (Oker-Bloom, 1989), as it is believed that IgG specific to the intact virion or its surface proteins are most relevant to protective immunity. Interactions of antibodies with the structurally dependent neutralizing domains of the virus leading to neutralization in the tissue culture well should not be directly extrapolated to be representative to protective immunity in the organism.  This concept was also  underlying the design of the experiments described in this thesis, where the biological function of purified proteins was assessed in order to prove that these proteins had maintained their native structure. On the basis of the presented data, it would be reasonable to speculate that more specific serological RV assay, correlating better with protective immunity, would entail the use of recombinant El, followed by  -106-  deglycosylation and cystine cleavage and blockage. A positive signal in such an assay will likely correlate well with functional T cell immunity and protection. Bacterially expressed RV proteins have been described and produced by independent groups (Londesborough et al., 1992; Dr. S.Gillam and H. Chaye, UBC, personal communication) and El has been expressed at high concentration in CHO cells (Dr.T.Hobman, University of California (San Diego), LaJolla, CA) and could potentially be employed in ELISA test using linearized RV El protein. The use of insect cell derived El, used in native conformation for anti-RV IgG ELISA, may be an expensive way of measuring irrelevant IgG and the use of synthetic peptides, cyclized or not, may be a costly and elaborate alternative for arriving at a product providing similar result as the linearized El ELISA.  -107-  SUMMARY AN]) PERSPECTIVES The hypothesis that intrauterine RV exposure leads to the development of RV-specific tolerance was tested by a comparative analysis of the RV protein-specific antibody responses in CRS and control patients. In order to examine the antibody responses of these patient groups a series of techniques had to be developed or optimalized. ELJSA techniques using whole RV were improved by including non-ionic detergent, resulting in disrupted virions and leading to increased sensitivity of the assay. The introduction of buffers, containing heat denatured blocking proteins, further improved the performance of RV EUSA as well as ELJSAs using other antigen systems (Mauracher et al., 1991b). For the determination of RV protein-specific antibody responses, both Western blot and ELJSA using separated proteins were developed.  Western blotting techniques were  developed using non-reducing conditions (Mauracher et al., 1992a; Zhang et al., 1992). This approach provided a better alternative than immunoprecipitation and allowed the detection of IgG, IgA and 1gM to the individual RV structural proteins.  Furthermore it showed that SDS-PAGE under low temperature and non-reducing  conditions did not selectively destroy the antigenicity of RV antigens.  Preparative SDS-PAGE followed by  electroelution was then employed to produce separated preparations of the individual RV structural proteins El, E2 and Cd. That proteins retained their antigenicity and much of their structural integrity was shown as both separated El and C were demonstrated to have maintained their biological functions. Optimal conditions for coating these proteins onto ELJSA microtitre plates were determined and pooled preparations of each protein were used to examine RV protein-specific IgG quantities and affmities. IgG affinity assays were initially evaluated using whole virus preparations.  Two techniques, the chaotropic  elution ELTSA and the one-well inhibition ELISA were examined in detail for the determination of relative average affinities of anti-RV IgG in human sera. The chaotropic elution EUSA using 8 M urea as a denaturant gave reproducible results and demonstrated affmity maturation of the anti-RV response following RV infection in adults. It was more applicable than the inhibition affinity assay as it was independent of RV-specific 1gM and required less antigen, which was an important consideration once affinity assays were used in separated RV protein EUSAs.  Biological activity of antibodies were determined by HAT, an assay which has long been  considered the standard to which other RV serology assays are compared. The separated RV proteins were analyzed for their structural integrity. This was felt to be important as the  -108-  majority of El and E2 epitopes had been shown to be structurally dependent (Mauracher et al., 1992b). As biological activity of proteins is generally dependent on structure, it was inferred that intact RV protein function would indicate structural integrity. The function of El protein is well defined and includes HA activity which has been localized to a defined region on the protein and is known to be sensitive to denaturation (Terry et al., 1988; Qui et al., 1992). The ability of separated El to compete for HAl IgG with native virions showed that the structurally dependent HA domain had remained intact in the El preparations.  During the investigations  of alternative methodologies for RV protein purification it was discovered that the C protein became detergentsoluble at pH below pH 5.5. This led to the further investigation of the solubility shift of this protein and resulted in the proposal of the intralysosomal uncoating mechanism of RV nucleocapsid (Mauracher et al., 1991a). This event is believed to be dependent on structural changes in the protein and has recently been shown to be applicable in SFV also, which has lead to the more broadly applicable proposal of intralysosomal capsid uncoating in the Togaviridae family (Schiegel and Kempf, 1992). Purified C protein was shown to have retained its biochemical property to undergo the low pH induced solubility change, indicating that at least one of this protein’s biological functions had remained intact throughout the separation protocol. It may be of interest to investigate compounds which may interfere with this property of the C protein, such as the WIN compounds (Badger et a!., 1988) which have been shown to inhibit the uncoating of picorna viruses. If low pH-induced uncoating is shown to be a common mechanism for RNA viruses which enter the cell by endocytosis such as Retroviridae or Flaviviridae, then such compounds may prove to become a viable alternative to anti-viral drugs based on nucleotide analogs. Once the immunological techniques had been established, it became possible to investigate the RV-specific antibody responses of adults with uncomplicated outcomes of Rubella in comparison with those of the CRS patient group. Healthy RV responders predominantly produced antibodies directed to the El protein which were of high affinity and appeared rapidly following infection. Of these, on average 90% were directed to conformationally-dependent epitopes with 10% being directed to linear epitopes. The E2-speciflc response was interesting in that it ethibited slow onset with IgG increasing in quantity as late as one year following infection. It remains unclear if this indicated that RV antigen may persist in the lymph or if crossreactivity between RV E2 and other antigens or self-constituents exists. The affinity of both anti-E2 and C IgG was observed to remain low throughout the first several years following Rubella infection, whereas El-specific IgG rapidly matured to -109-  high affinity in the months following Rubella. This demonstrated that differential affinity to individual proteins of a pathogen existed (Mauracher et a!., 1991) and indicated the importance of investigating the protein-specific immunity in the response to pathogens.  The immunodominant role of El in the antibody response was also  demonstrated in the lymphoproliferative response (Chaye et al., 1992), indicating that the El protein may be the most suitable candidate for the development of subunit vaccines. The comparative analysis of humoral responses to RV between CRS and normal control patients showed that the CRS patient group had significantly lowered IgG responses to the El protein, as measured on RV protein ELISA. These results supported fmdings by others (Katow and Sugiura, 1986). The most prominent feature which differentiated the IgG responses of CRS patients from those in control patients was the selective inability of the CRS group to produce IgG to linearized El. A model has been proposed which hypothesises that CRS patients have been selectively tolerized to the El protein of RV. This is reflected in serological responses in that IgG directed to linear epitopes is absent and the affmity of IgG to whole RV as well as El is low. Studies on the serological and cellular responses of CRS- and control patients to RV structural proteins suggest that the Th cell responses to El are significantly reduced in CRS patients (Chaye et al., 1992) and ongoing studies using peptides representing sequences of the RV structural proteins have provided further evidence that lymphoproliferative responses to El peptides are absent in most CRS patients whereas responses to E2 and C remain intact (Ou et al., 1992). This thesis has provided evidence to support the hypothesis that intrauterine RV exposure can lead to immunological tolerance.  Tolerance is proposed to be specific to the El protein and the absence of normal  cellular and serological responses may explain viral persistence in the CRS patient. Viral persistence may also be involved in the pathogenesis of RAA and it now is possible to determine if a similar deficit exists in the cellular responses of patients developing arthritis following RV infection. Models for tolerance induction in the human are rare and it is proposed that further studies on well defmed CRS patients will contribute to the understanding of the development of the human immune system. With the availability of peptides from all viral proteins, the fine specificity of both the B and T cell response can be studied. Genetic background of the CRS patients was not taken into account and will always remain a variable in human studies and may be a complicating factor in the detailed T cell studies of CRS and control patients. The selective tolerance to El remains intriguing and a good explanation why E2 and C responses are still -110-  present in the CRS patients has not been found. An answer to this question is important as it may teach us what characteristics make a membrane- or an internal protein good T cell antigens. Ultimately such answers may lead to more immunogenic vaccines or conversely they may provide us with clues how proteins can be treated to result in better graft acceptance.  -111-  REFERENCES Adkinson NF, Sabotka AK and Lichtenstein LM (1979): Evaluation of the quantity and affinity of human IgG blocking antibodies. J. Immunol. 122:965-972. Ahlstedt S, Holmgren J and Hanson LA (1974): Protective capacity of antibodies against E. coli 0 antigen with special reference to the avidity. mt. Arch. Allergy Appi. Immunol 46:470-480. Alford CA, Neva FA and Weller TH (1964): Virologic and serologic studies on human products of conception after maternal rubella. New. Engi. J. Med. 271:1275-1281. Allen D, Cumano A, Dildrop R, Kock SC, Rajewsky K, Rawjewsky N, Roes 3, Sablitzki F and Siekewitz M (1987): Timing, genetic requirements anf functional consequences of somatic hypermutation during B-cell development. Immunol. Rev. 96:5-22. Ai-Nakib W, Best 3M, Banatvala JE (1975): Rubella-specific serum and nasopharyngeal responses following naturally acquired and vaccine induced infection. Lancet 1:182-185. Alpers JH, Steward MW and Soothill IF (1972): Differences in immune elimination in inbred mice. The role of low affinity antibody. Clin. Exp. Immunol. 12:121-132. Anderson SG (1949): Experimental rubella in human volunteers. 3. Immunol. 62:29-40. Andersson B (1970): Studies on the regulation of avidity at the level of antibody forming cells. J. Exp. Med. 132:77-83. Avrameas S (1969): Coupling of enzymes to protein with gluteraldehyde. Use of the conjugates for the detection of antigen and antibodies. Immunochemistry 6:43-52. Badger J, Minor I, Kremer MJ, Oliveira MA, Smith TJ, Griffith IF, Guerin DMA, Krishnaswamy S, Luo M, Rossmami MG, McKinlay MA, Diana GD, Dutko FJ, Fancher M, Rueckert RR, Heinz BA (1988): Structural Analysis of a series of antiviral agents complexed with human rhinovirus 14. Proc. Nati. Acad. Sci. USA 85:33043308. Bardeletti G, Tektoff J, Lenches EM, Strauss EG and Strauss JH (1979): Rubella virus maturation and production in two host cell systems. Intervirology 11:97-103. Baric RS, Carlin U and Johnston RE (1983): Requirement for host transcription in the replication of sindbis virus. J. Virol 45:200-205. Barlow RM (1983): Some interactions of virus and maternal foetal immune mechanisms in Border disease of sheep. In: P0 Behan, V Ter Meulen and FC Rose (eds). “Immunology of Nervous System Infections”. Amsterdam: Elsevier. pp 255-268. Baron MD and Forsell K (1991): Oligomerization of the structural proteins of rubella virus. Virology 185:811 819.  -  Bellanti JA, Artenstein MS, Olsen LC, Buesher EL, Luhrs CE and Milstead KL (1965): Congenital rubella. Clinicopathologic, virologic and immunologic studies. Am. J. Dis. Child. 110:464-472. Berzofsky JA (1991): Mechanism of T cell recognition with application to vaccine design. Mol. Immunol. 28:217223. Bidwell DE, Bartlett A and Voller A (1977): Enzyme immunoassays for viral diseases. 3. Infect. Dis. 136:S274-  -112-  S278. Black FL (1966): Measles endemicity in insular populations: Critical community size and its evolutionary implications. J. Theoret. Biol. 11:207-211. Black FL, Woodall JP, Evans AS, Liebhaber H and Henle G (1970): Prevalence of antibody against viruses in the Tiriyo, an isolated Amazon tribe. Amer. 3. Epidemiol. 91: 430-438. Blank SE, Leslie GA and Clem LW (1972): Antibody affinity and valance in viral neutralization. J. Immunol. 108: 665-673. Bordier C (1981): Phase separation of integral membrane proteins in Triton X-1 14 solution. J. Biol. Chem. 256:1604-1607. Brush B (1989): “Characterization of monoclonal antibodies to the rubella virus structural proteins”. Vancouver, BC: University of British Columbia, M.Sc.Thesis Buchmeier MJ, Welsh RM, Dutko RJ and Oldstone MBA (1980): The virology and immunobiology of lymphocytic choriomeningitis virus infection. Adv. Immunol. 30:275-331. Buimovici-Klein E, O’Beime AJ, Millian SJ and Cooper LZ (1980): Low level rubella immunity detected by ELISA and specific lymphocyte transformation. Arch. Virol. 66:321-327. Buimovoci-Klein E, Weiss KE, Cooper LZ (1977): Interferon production in lymphocyte cultures after rubella infection in humans. J. Infect. Dis. 135:380-385. Buimovici-Klein E and Cooper LZ (1985): Cell-mediated immune response in rubella infections. Rev. Inf. Dis. 7:S123-S128. Burke DS, Nisalak A and Johnson DE (1988): A prospective study of dengue infection in Bankok. Am. J. Trop. Med. Hyg. 38: 172-180. Burnett FM and Fenner F (1949): “The Production of Antibodies”, 2nd ed, London, England: Macmillan. Carson LF, Twiggs LB, Fukushima M, Ostrow RW, Faras AJ and Okagaki T (1986): Human genital papilloma infections: an evaluation of the immunologic competence in the genital-neoplasia syndrome. Am. J. Obstet. Gynecol. 155:784-789. Center for Disease Control (1975): Reported morbidity and mortality in the united states. MMWR 24:2-4. Cemy A, Sutter S, Bazin H, Hengartner H and Zinkernagel RM (1988): Clearance of lymphocytic chorio meningitis virus in antibody- and B-cell deprived mice. J. Virol. 62:1803-1807. Chantler JK and Tingle AJ (1982): Isolation of rubella virus from human lymphocytes after acute natural infection. J. Infect. Dis. 145:673-677. Chantler 1K, Tingle AJ and Petty RE (1985): Persitent rubella virus infection associated with chronic arthritis in children. N. Engi. J. Med. 313:1117-1123. Chaouat G (1990): “The immunology of the fetus”. Boca Raton, Fl: CRC Press. Chase A (1984): “Magic shots: A human and scientific account of the long and continuing struggle to eradicate infectious diseases by vaccination”. New York, NY: William Morrow and Co. Chaye HH, Mauracher CA, Tingle AJ and Gillam S (1992): Cellular and humoral responses to rubella virus El, -113-  E2 and C proteins. J. Clin. Microbiol. (submitted) Chesebro B, Race R, Wehrly K (1985): Identification of scrapie prion protein-specific mRNA in scrapie-infected and uninfected brain. Nature 315:331-333. Clarke DM and Casals J (1958): Techniques for hemagglutination inhibition with arthropod borne viruses. Am. J. Trop. Med. and Hyg. 7:56 1-573. Clarke DM, Loo TW, Hui I, Chong P and Gillam S (1987): Nucleotide sequence and in vitro expression of rubella virus 24S subgenomic mRNA encoding the structural proteins El, E2 and C. Nuc. Acids Res. 15:30413057. Cochi SL, Edmons LE, Dyer K, Greaves WL, Marks JS, Rovira EZ, Preblud SR and Ortenstein WA (1985): Congenital rubella syndrome in the United States, 1970-1985. Am. J. Epidemiol. 129:349-361. Cooper MD and Butler IL (1989): Primary immunodeficiency diseases. In: WE Paul (Ed.) Immunology”, 2nd ed., New York: Raven Press, pp 1033-1058.  Fundamental  Cooper LZ and Krugman S (1967): Clinical manifestations of postnatal and congenital rubella. Arch. Ophthalmol. 77:434-439. Cooper LZ, Florman AL, Ziring PR and Krugman S (1971): Loss of rubella hemagglutinin inhibition antibody in congenital rubella, failure of seronegative children with congenital rubella to respond to HPV-77 vaccine. Am. J. Dis. Child. 122:397-403. Cooper LZ and Buimovici-Klein E (1985): Rubella. In Fields BN (ed): “Fields Virology” ,New York: Raven Press, pp 1005-1020. Coyle PK, Wolimsky JS, Buimovici-Klein E, Moucha R and Cooper (1982): Rubella-specific immune complexes after congenital infection and vaccination. Infect. Immunol. 36:498-503. Cunningham AL and Fraser IRE (1985): Persistent rubella virus infection of human synovial cells cultured in vitro. J. Infect. Dis. 151:638-645. Cusi MGH, Metelli R and Valensin PE (1989): Immune responses to wild and vaccine rubella viruses after rubella vaccination. Arch. Virol. 106:63-72. Daffos F, Grangeot-Keros L, Lebon P, Forestier F, Pavlovsky MC and Pillot J (1984): Prenatal diagnosis of congenital rubella. Lancet 2:1-3. Davies WJ, Larson HE, Simserian JP, Parkman PD and Meyer HM (1971): A study of rubella immunity and resistance to infection. JAMA 215:600-608. Decker IL, Malone DG and Boulos H (1984): Rheumatoid arthritis: Evolving concepts of pathogenesis and therapy. Ann. Intern. Med. 101:810. de Mazancourt A, Watham MN, Nicholas JC and Wolinsky JS (1986): Antibody response to the rubella virus structural proteins in infants with the congenital rubella syndrome. I. Med. Virol. 19:111-122. Desmond MM, Wilson GS and Melnick IL (1967): Congenital rubella encephalitis: course and early sequelae. J. Pediatr. 71:311-331. Dimech W, Bettoli AB, Eckert D, Francis B, Hamblin J, Kerr T, Ryan C and Skurrie I (1992): Multicentre evaluation of five commercial rubella virus immunoglobulin G kits which report in international units per milliliter. J. Chin. Microbiol. 30:633-641. -114-  Dorsett PH, Miller DC, Green KY and Byrd Fl (1985): Structure and function of the rubella virus proteins. Rev. Inf. Dis. 7:S150-S156. Eggerding FA, Peters J, Lee RK and Inderlied CB (1991): Detection of rubella virus gene sequences by enzymatic amplification and direct sequencing of amplified DNA. J. Clin. Microbiol. 29:945-952. Eisen HN and Karush F (1949): The interaction of purified antibody with homologous hapten. Antibody valance and binding constant. 3. Am. Chem. Soc. 71:363-370. Eisen HN and Siskind GW (1964): Variations in affmities of antibodies during the immune response. Biochemistry 3:996-1008. Enders G and Jonatha W (1987): Prenatal diagnosis of interuterine rubella. Infection 15:162-164. Enders G, Miller E, Nickerl-Pacher U and Cradock-Watson JE (1988): Outcome of confirmed pericon-ceptional maternal rubella. Lancet 1:1445-1447. Engvall E and Perlmann P (1971): Enzyme-linked immunosorbent assay (ELISA). Quantitative assay of IgG. Immunochem. 8:871-874. Fauci AS, Frank MM and Johnson JS (1970): The relationship between antibody affinity and efficiency of complement fixation. J. Immunol. 105:215-220. Fenner F, McAuslan BR, Mims CA, Sambrook J and White DO (1974): “The Biology of Animal Viruses’. New York: Academic Press. Field AM, Vandervelde EM, Thompson KIvI and Hutchinson DN (1967): A comparison of the hemagglutination inhibition test and the neutralization test for the detection of rubella antibody. Lancet 11:182-184. Fitzgerald MG, Pullen GR and Hosking CS (1988): Low affmity antibody to rubella antigen in patients after rubella infection in utero. Pediatrics 81:812-824. Fitzpatrick DR and Bielefeldt-Ohmann H (1991): Mechanism of herpesvirus immuno-evasion. Microbial Pathogenesis 10:253-259. Forbes JA (1969): Rubella: Historical aspects. Am. J. Dis. Child. 118:5-11. Ford DK, Tingle AJ and Chantler 1K (1986): Rubella arthritis. In L Espinoza, DL Goldenberg, FC Arnett and GS Alceron (eds): “Infections in the rheumatic diseases”. New York, NY: Grune & Stratton Inc., pp 103-106. Frey TK, Marr ID, Hemphill ML and Domingues G (1986): Molecular cloning and sequencing of the region of rubella virus genome coding for glycoprotein El. Virology 154:228-232. Friguet B, Djavadi-Ohaniance L and Goldberg ME (1984): Some monoclonal antibodies raised with native protein bind preferentially to the denatured form. Mo!. Immunol. 21:673-677. Froshauer 5, Kartenbeck J, and Helenius A (1988): Alphavirus RNA replicase is located on the cytoplasmic surface of endosomes and lysosomes. 3. Cell Biol. 107:2075-2080. Fuccillo DA, Steele RW, Heusen SA, Vincent MM, Hardy TB and Bellanti JA (1973): Impaired cellular immunity to rubella virus in congenital rubella. Infect. Immun. 8:81-83. Fujinami RS and Oldstone MBA (1984): Antibody initiates virus persistence: Immune modulation and measles virus infection. In AL Notkins and MBA Oldstone (eds): “Concepts in Viral pathogenesis”. New York, NY: Springer Verlag. pp 187-193. -115-  Fuller SD and Argos P (1987): Is sindbis a simple picornavirus with an envelope? EMBO J. 6:1099-1105. Goodnow CC, Adeistein S and Basten A (1990): The need for central and periferal tolerance in the B cell repertoire. Science 248:1373-1379. Green KY and Dorsett PH (1986): Rubella virus antigens: Localization of epitopes involved in hemagglutination and neutralization by using monoclonal antibodies. J. Virol. 57:893-898. Green RH, Balsamo MR, Giles JP, Krugman S and Mirick GS (1965): Studies on the natural history and prevention of rubella. Am. J. Dis. Child. 110:348-365. Gregg NM (1941): Congenital cataracts following german measles in the mother. Trans. Ophthalmol. Soc. Aust. 3 :35-46. Hamilton RG (1987): “The Human IgG Subclasses”. San Diego: Hoechst Celanese Corp. p 32. Hancock EJ, Pot K, Puterman ML and Tingle AJ (1986): Lack of association between titres of HAl antibody and whole-virus ELISA values for patients with congenital rubella syndrome. J. Inf. Dis. 154:1031-1033. Hardy JB (1970): Response to rubella vaccination among seronegative children with congenital rubella. Pediat. Res. 4:513-514. Helenius A, Marsh M and White J (1982): Inhibition of Semliki Forest virus penetration by lysosomotrophic weak bases. J. Gen. Virol. 58:47-61. Hemphill ML, Forng R, Abernathty ES and Frey TK (1988):Time course of virus-specific macromolecular synthesis during rubella virus infection in Vero cells. Virology 162:65-75. Hedman K, Hietala J, Tiilikainen A, Hartikainen-Sorri A-L, Ráihá, Suni J, Vánnen P, Pietiláinen M (1989): Maturation of imniunoglobulin G avidity after rubella vaccination studied by enzyme linked immunosorbent assay (Avidity-ELISA) and by haemolysis typing. 3. Med. Virol. 27:293-298. Hedman K and Rousseau SA (1989): Measurement of avidity of specific IgG in the verification of recent primary rubella. J. Med. Virol. 27:288-292. Heggie AD and Robbins FC (1969): Natural rubella aquired after birth: clinical features and complications. Am. J. Dis. Child. 118:12-17. Herrmann KL (1985): Available rubella serologic tests. Rev. Infect. Dis. 7:S108-S1 12. Hinman AR, Bart 10, Orenstein WA and Preblud SR (1983): Rational strategy for rubella vaccination. Lancet 1: 39-41. Hirano N, Kao M and Ludwig H (1983): Persitant, tolerant or subacute infection in Borna disease virus-infected rats. 3. Gen. Virol. 64: 1521-1530 Hiro Y and Tasaka S (1938): German measles is a virus disease. Monatschr. Kinderk. 76:328-332. Hirsch MS and Curran 3 (1990): Human immunodeficiency viruses, biology and medical aspects. In BN Fields and DM Knipe (eds): “Fields Virology”. New York, NY: Raven Press, ppl545-l57O. Hobman TC, Lundstr6m ML and Gillam S (1990): Processing and intracellular transport of rubella virus structural proteins in COS cells. Virology 178:122-133. Hobman TC (1989) “Molecular Cell Biology of Rubella Virus Structural Proteins.” Ph.D. Thesis, University of -116-  British Columbia, Vancouver B.C. Hobman TC and Gillam S (1989): In vitro and in vivo expression of rubella virus glycoprotein E2: The signal sequence is contained in the C-terminal region of the capsid protein. Virology 173:241-250. Hobgen DN, Brown SE, Howard CR and Steward MW (1986): HSsAg:anti-HBs immune complexes. A method for separating the constituent components and assessment of affinity of the antibody. 3. Immunol. Methods 93:29-36. Horzinek MC (1981): “Non-Arthropod-Borne Togaviruses”. London: Academic Press. Ho-Terry L and Cohen A (1982): Rubella virion polypeptides: Characterization electrophoresis, isoelectric focusing and peptide mapping. Arch. Virol. 72:47-54.  by polyacrylamide gel  Ho-Terry L and Cohen A (1984): The role of glycosylation on haemagglutination and immunological reactivity of rubella virus. Arch. Virol. 79: 139-146. Ho-Terry L and Cohen A (1985): Rubella virus haemagglutinin: Association with a single virion glycoprotein. Arch. Virol. 84:207-215. Hovi T and Vaheri A (1970): Infectivity and some physiochemical characteristics of rubella virus ribonucleic acids. Virology 42:1-8. Howson CP, Howe CJ and Fineberg HV (1991): “Adverse Effects of Pertussis and Rubella Vaccines: A Report of the Committee to Review the Adverse Cosequences of Pertussis and Rubella Vaccines”. Washington,DC: National Academy Press. Howson CP and Fineberg HV (1992): The recochet of magic bullets: Summary of the institute of medicine report, adverse effests of pertussiss and rubella vaccines. Pediatrics 89:318-324. Huang AS (1988): Modulation of viral disease processes by defective interfering particles. In: E Domingo, JJ Holland, P Ahiquist (eds): “RNA Genetics III”. Boca Raton, FL: CRS Press, pp 195-208. Hutchings DL, Van Drunen Little-Van den Hurk S and Babiuk LA (1990): Lymphocyte proliferative responses to separated bovine herpes virus I protein in immune cattle. J. Virol. 64:5 1 14-5 122. Ilonen J and Salmi A (1986): Comparison of HLA-Dwl and -Dw2 positive adherent cells in antigen presentation to heterozygous T-cell lines: a low rubella antigen-specific response association with HLA-Dw2. Hum. Immunol. 17:94-101. Inouye 5, Hasegawa A, Matsuno S and Katow S (1984): Changes in antibody avidity after virus infections: detection by an immunosorbent assay in which a mild protein-denaturing is employed. 20:525-529. Inman RD, Chiu B and Hamilton NC (1987): Analysis of immune complexes in rheumatoid arthritis for EpsteinBarr virus antigens reveals cross-reactivity of viral capsid antigens and human IgG. J. Immunol. 138:407-412. Jerne NK (1974): Toward a network theory of the immune system. Ann. Immunol. 125c:373-389. Johausson BE, Moran TM and Kilboume ED (1987): Antigen-presenting B-cells and helper T-cells cooperatively mediate intravironic antigenic competition between influenza A virus surface glycoproteins. Immunology 84:68696873. Kalkkinen N, Oker-Blom C and Pettersson RF (1984): Three genes code for rubella virus structural proteins El, E2a, E2b and C. J. Gen. Virol. 65: 1549-1557.  -117-  Kamoun PP (1988): Denaturation of globular proteins by urea: breakdown of hydrogen or hydrophobic bonds? TIBS 13:424-425. Katow S and Sugiura A (1985): Antibody response to individual rubella virus protein in congenital and other rubella virus infections. J. Clin. Microbiol 21:449-451. Katow S and Sugiura A (1988): Low pH-induced conformational changes of rubella virus envelope proteins. J. Gen. Virol. 69:2797-2807. Kawano K and Minamishima Y (1987): Removal of nonspecific hemagglutination inhibitors, immunoglobulin G, and immunoglobulin A with streptococcal cells and its application to the rubella hemagglutination inhibition test. Arch. Virol. 95:41-52. Kempf C, Michel MR, Kohier U and Koblet H (1987): Can viral envelope proteins act as or induce proton channels? Biosci. Rep. 7: 76 1-769. Kielian M and Helenius A (1988): Entry of aiphaviruses. In S Schlesinger and MJ Schlesinger (eds): “The Togaviridae and Flaviviridae” .2nd ed., New York, NY: Plenum Press, pp 91-119. Kilroy AW, Schaffner W, Fleet WF, Lefkowitz LB, Karzon DT and Fenichel GM (1970): Two syndromes followinf rubella immunization: clinical observations and epidemiological studies. JAMA 214:2287-2292. Kienk H-D (1990): Influence of glycosylation on antigenicity of viral proteins. In MHV Van Regenmortel and AR Neurath (eds): “Immunochemistry of Viruses II”. Elsevier: Amsterdam, pp 25-37. Koch S (1989): The neonatal human’s immune response to herpes simplex virus infection: a critical review. Pediatr. Infect. Dis. 8:67-74. Laemmli UK (1970): Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature (London) 227:680-685. Leinikki P0, Shekarchi I, Dorsett P abd Sever JL (1978): Enzyme-linked immunosorbent assay determination of specific rubella antibody levels in micrograms of immunoglobulin G per milliliter of serum in clinical samples. 3. Clin. Microbiol. 8:419-423. Lennette EH, Schmidt NJ and Magoffm R (1967): The hemagglutination inhibition test for rubella: a comparison of its sensitivity to that of neutralization, complement fixation and flourescent antibody test for diagnosis of infection and determination of immune status. J. Immunol. 99:785-793. Lennette EH, Jung M and Jung F (1989): “Serologyof Infectious Disease Syndromes” Cham, Switzerland: Virion Edition. Levine BB and Levytska V (1967): A sensitive hemagglutination assay method for dinitrophenyl-specific antibodies. The effect of antibody binding affinity on titres. J. Immunol. 98:648-652. Levinson SS and Goldman J (1988): Interference by endogenous immunoglobulins with ligand-binding assay. Clin. Immunol. Newslett. 9:101. Liu Y-J, Johnson GD, Gordon 3 and MacLennnan 1CM (1992): Germinal centres in T-cell-dependent antibody responses. Immunol. Today 13:17-21. Londesborough P, Terry G and Ho-Terry L (1992): Reactivity of a recombinant rubella El antigen expressed in E.coli. Arch. Virol. 122:391-397.  -118-  Loo TW, MacDonald I, Clarke DW, Trudel M, Tingle AJ and Gillam S (1986): Detection of antibodies to individual proteins of rubella virus. J. Virol. Meth. 13:149-159. Lund K and Chantler JK (1991): Molecular mimicry between rubella virus and a component of human synovial cells. The American Society for Virology, 1991 annual meeting, Ft. Collin, Co, USA, July 7th-i ith 1991. Lundstr6m ML, Mauracher CA and Tingle AJ (1991): Characterization of carbohydrates linked to rubella virus glycoprotein E2. J. Gen. Virol. 72: 843-850. McChesney MB and Oldstone MBA (1987): Viruses perturb lymphocyte functions: Selected principles characterizing virus-induced immunosuppression. Ann. Rev. Immunol. 5:279-304. McDonald H (1991): “Studies on the processing of rubella virus structural proteins by analysis of the endoproteolytic cleavage sites”. Vancouver, BC: University of British Columbia, M.Sc. Thesis McDonald H, Hobman TC and Gillam S (1991): The influence of capsid protein cleavage on the processing of E2 and El glycoproteins of rubella virus. Virology 183:52-60. Mcllvane (1962): Buffer Solutions. In K Diem (ed): “Documenta Geigy, Scientific Tables”, 6ed, Geigy Pharmaceuticals: Montreal, pp 314-315. McKinley MA, Frank JA (1986): Win 51711- A novel drug for the treatment of enterovirus infections. In J Mills and C Lawrence (eds): “Antiviral Chemotherapy, New Directions for Clinical Applications and Research”. Amsterdam: Elsevier, pp 90-96. McNabb T, Koh TY, Dorrington 1<3 and Painter RH (1976): Structure and function of the inimunoglobulin domains. V. Binding of immunoglobulin G and fragments to placental membrane preparations. J. Immunol. 117:882. Maatsuura Y, Possee RD, Overton HA and Bishop DHL (1987): Baculovirus expression vectors: The requirement for high level expression of proteins, including glycoproteins. J. Gen. Virol. 68:1233-1250. Maes R, Vaheri A, Sedgwick WD and Plotkin SA (1966): Synthesis of virus and macromolecules by rubellainfected cells. Nature 210:384-385. Malancon P and Garoff H (1987): Processing of the semliki forest virus structural polyprotein: role of the capsid protease. J. Virol. 61:1301-1309. Maniatis T, Fritsch EF and Sambrook 3 (1982): “Molecular Cloning: a laboratory manual”. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Publications, pp 458-463. Manser T (1990): The efficiency of antibody affinity maturation: Can the rate of B-cell division be limiting. Immunol. Today 11:305-308. Marcy SM and Kibrick 5 (1972): Exanthems. In Hoeprich PD (ed): “Infectious Diseases” Hagerstown, MA: Harper and Row, pp 875-879. Marrack P and Kappler J (1990): The staphylococcal enterotoxins and their relatives. Science 248:705-711. Mastromarino P, Rieti 5, Pozzi D, Grimaldi S and Orsi N (1990): Fusion of rubella virus with vero cells. Abstracts of the Vifith mt. Cong. Virol. 388. Mauracher CA, Gillam 5, Shukin R and Tingle AJ (1991a): pH dependent shift of rubella virus capsid protein. Virology 181:773-777.  -119-  Mauracher CA, Mitchell LA and Tingle AJ (1991b): Reduction of rubella EUSA background using heat denatured sample buffer. I. Immunol. Meth. 145:251-254. Mauracher CA, Mitchell LA and Tingle AJ (1992a): Differential antibody affinity to the structural proteins of rubella virus. J. Med. Virol. 36:202-208. Mauracher CA, Mitchell LA, Ddcarie D, Ho KL, Shukin R, Zrein M, Lacroix M and Tingle AJ (1992b): Selective tolerance to rubella virus structural protein in congenital rubella syndrome (CRS): A potential mechanism for viral persistence. Canadian Society of Immunology Mauracher CA, Mitchell LA and Tingle AJ (1992c): El specific tolerance in CRS patients. Manuscript in preparation. Meurman 0, Waris M and Hedman K: IgG antibody affinity in respiratory syncytial virus infection. Manuscript in preparation. Mikki NPH and Chantler JK (1992): Differential ability if wild-type and vaccine strains of rubella virus to replicate and persist in human joint tissue. Chin. Exp. Rheumatol. 10:2-12. Miller E, Cradock-Watson JE and Pohlock TM (1982): Consequences of confirmed maternal rubella at successive stages of pregnancy. Lancet 2:781-784 Milstein C (1987): Diversity and the generation of high affinity antibodies. Biochem. Soc. Trans. 15:779-787. Mitchell LA, Zhang T, Mauracher CA and Tingle AJ (1992a): Differential antibody response to rubella virus infection in male and female patients. In press, J. Clin. Microbiol. Mitchell LA, Zhang T, Ho M, Ddcarie, Tingle AJ, Zrein M and Lacroix M (1992b): Characterization of rubellaspecific antibody responses using a new synthetic peptide-based enzyme-linked immunosorbent assay. In press, J. Med. Microbiol. Morris GE, Coleman RM, Best JM, Benetato BB and Nahmias Al (1985): Persistence of serum IgA antibodies to herpes simplex, varicehla zoster, cytomegalovirus and rubella virus detected by enzyme-linked immunosorbent assays. 3. Med. Virol. 16:343-349. Munro ND, Smithhells RW, Sheppard S, Holzel H and Jones G (1987): Temporal relations between maternal rubella and congenital defects. Lancet 2:201-204. Murphy PA, Halonen PE and Harrison AK (1968): Electronmicroscopy of the development of rubella virus. J. Virol. 2:1223-1227. Nakhasi HL, Thomas D, Zheng D and Liu T-Y (1989a): Nucleotide sequence of capsid, E2 and El genes of rubella virus vaccine strain RA 27/3. Nucl. Acids Res. 17: 4393-4394. Nakhasi HL, Zheng D, Callahan L, Dave JR and Liu T-Y (1989): Rubella virus: mechanism of attenuation in the vaccine strain (HPV77). Virus Research 13:231-244. Narayan 0, Zink MC and Huso D (1988): Lentiviruses of animals are biological models of the human immunodeficiency viruses. Microbial Pathogenesis 5:149-157. Nash A (1985): Tolerance and supression in virus disease. Brit. Med. Bulletin 41:41-45, Nawa M (1979): Focus formation by rubella virus in rabbit kidney (RX-13) cell cultures and its application to the virus’ titration. Arch. Virol. 60:75-78.  -120-  Nemazee D and Bärki K (1989): Clonal deletion of autoreactive B lymphocytes in bone marrow chimeras. Proc. Natl. Acad. Sci. USA. 86:8039-8043. Neuhoff V, Arold N, Taube D and Ehrhardt W (1988): Improved staining of proteins in polyacrylamide gels including isoelectñc focusing gels with clear background at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250. Electrophoresis 9:255-262. Noelle NJ and Snow CE (1990): Cognate interaction between helper T cells and B cells. Immunol. Today 11:361368. Nossal JV (1991): B-cell selection and tolerance. Current Opinion in Immunol. 3:193-198. Ogra PL, Ogra SS, Chiba Y, Dzierba IL and Herd 1K (1975): Rubella virus infection in juvenile rheumatoid arthritis. Lancet 1:1157-1161. Oker-Blom C, Kalkkinen N, Káiriáinen L and Pettersson RF (1983): Rubella virus contains one capsid protein and three envelope glycoproteins, El, E2a and E2b. 3. Virol. 46:964-973. Oker-Blom C (1984a): The gene order for rubella virus structural proteins is N112-C-E2-El-COOH. J. Virol. 51:354-358. Oker-Blom C, Ulmanen I, Kaarianen L and Petterson RF (1984b): Rubella virus 40S RNA specifies a 24 S subgenomic RNA that codes for a precursor to structural proteins. J. Virol. 49:403-408. Oker-Blom C, Pettersson RF and Summers MD (1989): Baculovirus polyhedrin promoter-directed expression of rubella virus envelope glycoprotein El and E2, in Spodoptera frugiperda cells. Virology 172:82-91. Oldstone MBA and Dixon FJ (1969): Pathogenesis of chronic disease associated with persitent lymphocytic choriomeningitis viral infection. I. Relashionship of antibody productionto disease in neonatally infected mice. J. Exp. Med. 129:483-505. Oldstone MBA (1989): Viral persistence. Cell 56:517-520. Oshiro LS, Schmidt NJ and Lennette EH (1969): Electron microscopic studies of rubella virus. J. Gen. Virol. 5:205-210. Ou D, Chong P, Tripet B, Jefferies WA and Gillam 5: Characterization of the specificity and genetic restriction of human CD4+ T cell clones reactive to capsid antigen in rubella virus. Submitted to Virology. Paliard X, Sterling G, West 0, Lafferty JA, Clements IR, Kappler JW, Marrack P and Kotzin BL (1991): Evidence for the effects of a superantigen in rheumatoid arthritis. Science 253:325-329. Papadogiannakis N, Johnsen S-A and Olding LB (1985): Strong prostaglandin associated suppression of the proliferation of human matemal lymphocytes by neonatal lymphocytes linked to T versus B cell interactions and differential PGE 2 sensitivity. Clin. Exp. Immunol. 61:125-134. Parkman PD, Buesher EL and Artenstein MS (1962): Recovery of rubella virus from army recruits. Proc. Soc. Exp. Biol. Med. 111:225-230. Parkman PD, Buescher EL, Artenstein MS, McCown 3M, Mundon FK, and Druzt AD (1964): Studies of Rubella. I. Properties of the virus. 3. Immunol. 93:595-607. Partanen P, Seppãnen H, Suni 3 and Vaheri A (1985): Selective reactivity of antibodies to human immunoglobul ins 0, M and A with rubella virus proteins. 3. Clin. Microbiol. 21:800-802.  -121-  Payment P, Ajdukovic D, and Pavilanis V (1975): Le virus de la rubeole. I. Morphologie et proteines structurales. Can. 3. Microbiol. 21:703-709. Plotkin SA and Vaheri A (1967): Human fibroblasts infected with rubella virus produce a growth inhibitor. Science 156:659-661. Plotkin SA, Farquhar JD and Ogra PL (1973): Immunological properties of RA27/3 rubella virus vaccine. JAMA 225:585-590. Polk BF, Modlin JF, White JA and de Girolami PC (1982): A controlled comparison of joint reactions among women receiving one of two rubella vaccines. Am. J. Epidem. 115:19-25. Porterfield JS, Casals J, Chumakov MP, Gaidamovich SY, Hannoun C, Holmes 11-I, Horzinek M, Mussgay M, Oker-Blom N, Russel PK and Trent DW (1978): The Togaviridae. Intervirology. 9:129-148. Preblud SR (1986): Some current issues relating to rubella vaccine. JAMA 254:253-256. Qiu Z and Gillam S: The influence of N-linked glycosylation on the antigenicity and immunogenicity of rubella virus El glycoprotein. Submitted to Virology. Rabinowe SL, George KL, Loughlin R and Eisenbarth GS (1986): Congenital rubella, monoclonal antibodydefined T cell abnormalities in young adults. Am. J. Med. 81:779-783. Rath 5, Stanley CM and Steward MW (1988): An inhibition enzyme immunoassay for estimating relative antibody affinity and affinity heterogeneity. J. Immunol. Meth. 106:245-249. Rawis WE and Melnick JL (1966): Rubella virus carrier cultures derived from congenitally infected infants. J. Exp. Med. 123:795-816. Robinson RG, Dudenhoeffer FE and Hoiroyd HJ (1982): Rubella immunity in older teenagers and young adults: a comparison of immunity in those previously immunized with those unimmunized. J. Pediatr. 101:188-191. Ruscher IR, Javaherian K and McDonald C (1988): Antibodies that inhibit fusion of human iinmunodeficiency virus-infected cells bind a 24 amino acid sequence of the viral envelope gpl2O. Proc. Nati. Acad. Sd. USA. 85:3198-2202. Saiki RK, Gelfand DII, Stoffel 5, Scharf SH, Higuchi R, Horn GT, Mullis KB and Erlich HA (1988): Primerdirected enzymatic amplification od DNA with thermostable DNA polymarase. Science 239:487-491. Salonen E-M, Hovi T, Meurman 0, Vesikari T and Vaheri A (1985): Kinetics of specific IgA, IgD, IgE, IgG and 1gM antibody responses in rubella. 3. Med. Virol. 16:1-9. Sarnesto 5, Ranta 5, Vaananen P and Makela 0 (1985): Proportions of Ig classes and subclasses in rubella antibodies. Scand. 3. Inimunol. 21:275-283. Sato H, Albrecht P, Raynolds W, Stagno S and Ennis FA (1979): Transfer of measles, mumps and rubella antibodies from mother to infant. Am. J. Dis. Child. 133:1240-1243. Saulsbury PT, Winkeistein JA and Yolken RH (1980): Chronic rota virus infection in immunodeficiency. J. Pediatr. 97:61-65. Sauters SVH and Utsinger PD (1978): Viral arthritid. In FR Schmid (ed): “Clinics in Rheumatic Disease”. London: W.B. Saunders Company, pp225-240. Seppánen H, Huhtala M-L, Vaheri A, Summers MD and Oker-Blom C (1991): Diagnostic potential of -122-  baculovirus-expressed rubella virus envelope proteins. J. Clin. Virol. 29:1877-1882. Seto NOL, Hobman TC and Gillam S (1991): Expression and secretion of soluble form of rubella virus glycoprotein El and E2 in mammalian and insect cells. The American Society of Virology, 1991 annual meeting, Ft. Collins, Co, USA, July 7th-llth 1991. Scherle PA and Gerhard W (1988): Differential ability of B cells specific for external vs. internal influenza proteins to respond to help from influenza virus-specific T-cells in vivo.Proc. Nati. Acad. Sci. USA 85:4446-4450. Schlegel A and Kempf C (in press): A viral proton channel. In JAF Op den Kamp (ed): “Dynamics of Membrane Assembly”. NATO ASI. Schlesinger MJ (1987): The replication of togaviridae and flaviviridae at the molecular level. In RP Bercoff (ed): “The Molecular Basis of Viral Replication”. New York: Plenum Press, pp 217-238. Schlesinger S and Schlesinger MJ (1990): Replication of togaviridae and flaviviridae. In Fields B and Knipe DM (eds): “Fields Virology”,New York, NY: Raven Press, pp 697-711. Sharon N and Lis H (1982): Glycoproteins. In, H Neurath and RL Hilim (eds): “The Protein”, vol 5, 3rd ed., New York: Academic Press, pp 1-114. Shekarchi IC, Sever IL, Tzan N, Ley A, Ward LC and Madden D (1981): Comparison of hemagglutination inhibition test and enzyme-linked imrnunosorbent assay for determining antibody to rubella virus. J. Clin. Microbiol. 13:850-854. Shirley JA, Revill 5, Cohen BJ and Buckley MM (1987): Serological study of rubella-like illnesses. J. Med. Virol. 21:369-379. Singh I and Helenius A (1991): Ribosomes mediate uncoating of semliki forest virus nucleocapsid in the cytoplasm of the cell in vitro. Annual Meeting The American Society for Virology, Fort Collins, CO, July 7. -11. ,  Singh VK, Tingle AJ, and Schulzer M (1986): Rubella-associated arthritis. II. Relationship between circulating immune complex levels and joint manifestations. Ann. Rheum. Dis. 45:115-119. Six HR, Uemera K and Kinsky SC (1973): Effects of immunoglobulin class and affmity on the initiation of complement dependent damage to liposome model membranes sensitized with dinitrophenylated phospholipids. Biochemistry 12:4003-4011. Smith CA, Petty RE and Tingle AJ (1987): Rubella virus and arthritis. Rheum. Dis. Clin. North Am. 13:265-274. South MA and Sever JL (1988): Teratogen update: the congenital rubella syndrome. Teratology 31:297-307. Spruance SL, Mock LE, Bailey A, Ward JR and Smith CB (1972): Recurrent joint symptoms in children vaccinated with HPV-77DK12 rubella vaccine. J. Pediatr. 80:413-417. Steckelberg 3M and Cockerill ifi FR (1988): Serologic testing for human immunodeficiency virus antibodies. Mayo Chin. Proc. 63:373-380. Steece RS, Talley MS, Skeels MR and Lanier GA (1984): Problems in determining immune status in borderline specimens in an enzyme imrnunoassay for rubella immunoglobulin G antibody. J. Clin. Microbiol. 19:932-925. Steinman RM (1991): The dendritic cell system and its role in imrnunogenicity. Ann. Rev. Immunol. 9:271-296. Steward MW and Petty RE (1972): The antigen-binding characteristics of antibody pools of different relative affinity. Immunology 23.881-887. -123-  Steward MW and Steensgaard (1983): “Antibody affmity: Thermodynamic aspects and biological significance”, Boca Raton, Fl: CRC Press. Steward MW and Lew AM (1985): The importance of antibody affinity in the performance of immunoassays for antibody. J. Immunol. Methods 78:173-190. Steward MW, Stanley CM, Dimarchi R, Mulcahy G and Doel TR (1991): High-affinity antibody induced by immunization with a synthetic peptide is associated with protection of cattle against foot-and-mouth disease. Immunol. 72:99-103. Stewart GL, Parkman PD, Hopps HE, Douglas RD, Hamilton JP and Meyer HM Jr. (1967): Rubella-virus hemagglutination-inhibition test. N. Engi. J. Med. 276: 554-557. Stokes A, Mims CA and Grahame R (1986): Subclass distribution of IgG and IgA responses to rubella virus in man. J. Med. Microbiol. 21:283-285. Suomalainen M, Garoff H and Baron M (1990): The E2 signal sequence of rubella remains part of the capsid protein and confers membrane association in vitro. J. Virol. 64:5500-5509. Szakal AK, Kosco MR and Tew JG (1989): Microanatomy of lymphoid tissue during humoral immune responses: structure function relationships. Annu. Rev. Immunol. 7:91-109. Tanaka K, Isselbacher KJ, Khoury G and Jay G (1985): Reversal of oncogenesis by the expression of a major histocompatibility complex class I gene. Science 228:26-30. Tardieu M, Grospierre B, Durandy A and Griscelli C (1980): Circulating immune complexes containing rubella antigens in late-onset rubella syndrome. J. Pediatr. 97:370-373. Terry GM, Ho-Terry L, Londesborough P and Rees KR (1988): Localization of the rubella El epitopes. Arch. Virol. 98: 189-197. Theopold U and Kóliler G (1990): Partial tolerance in b-galactosidase-transgenic mice. Eur. J. Immunol. 20:1311-1316. Thomas MIJ and Morgan-Capner P (1988): Specific IgG subclass antibody in rubella virus infections. Epidemiol. Infect. 100:443-454. Thomson AR and Marker 0 (1989): MHC and non-MHC genes regulate elimination of lymphocytic chorio meningitis virus and antiviral cytotoxic T lymphocyte and delayed-type hypersensitivity mediating T lymphocyte activity in parallel. J. Immunol. 142:1333-1341. Tilley P and Junker AK: Low avidity anti-varicella zoster virus antibody responses in children with recurrent episodes of chickenpox. Manuscript in preparation. Tingle AJ,Allen M, Petty RE, Kettyls GD and Chantler 1K (1986): Rubella-associated arthritis. I. Comparative study ofjoint manifestations associated with natural rubella infection and RA27/3 rubella immunization. Ann. Rheum. Dis. 45:110-114. Tingle AJ (1990): Evidence linking rubella vaccine to chronic arthritis. Presented at the Institute of Medicine Workshop on Adverse Consequences of Pertussis and Rubella Vaccines, Washington, DC. Tondury G and Smith DW (1966): Fetal rubella pathology. J. Pediatr. 58:867-879. Tony H-P and Parker C (1983): Major histocompatibility complex-restricted polyclonal B cell responses resulting from helper T cell recognition of anti immunoglobulin presented by small B lymphocytes. J. Exp. Med. 161:223-124-  231. Towbin H, Stálilin T and Gordon J (1979): Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheats: Procedure and some applications. Proc. Nati. Acad. Sci. USA 76:4350-4354. Traub E (1938): Factors influencing the persistence of choriomeningitis virus in the blood of mice after clinical recovery. J. Exp. Med. 68:229-250. Udea K, Nishida Y, Oshima K, Yoshikawa H, Ohashi K and Nonaka 5 (1975): Seven-year follow-up study of rubella syndrome in Ryukyu with special reference to persistence of rubella hemagglutination inhibition antibodies. Jap. J. Microbiol. 19: 181-185. United States Department of Health, Education and Welfare (1970): CDC standard rubella hemagglutination inhibition test. Center for Disease Control of the Public Health Service, Atlanta GA. Vannanen P and Kaariainen L. Fusion and hemolysis of erythrocytes caused by three togaviruses: Semliki Forest, sindbis and rubella. J. Gen. Virol. 46:467-475. Van Regenmortel MHV (1990): The structure of viral epitopes. In MHV Van Regemnortel and AR Neurath (eds): “Immunochemistry of Viruses II”. Amsterdam: Elsevier, pp 1-17. Vaheri A and Hovj T (1972): Structural proteins and subunits of rubella virus. J. Virol. 9:10-16. Veale H (1866): History of an epidemic of rötheln with observations on its pathology. Edinburgh Med. 3. 12:404414. Vidgren G, Taldcinen K, Kalkkinen N, Kdridinen L and Pettersson RF (1987): Nucleotide sequence of the genes coding for the membrane glycoproteins El and E2 of rubella virus. J. Gen. Virol. 68:2347-2357. von Bonsdorff C-H and Vaheri A (1969): Growth of rubella virus in BHK21 cells: Electron microscopy of morphogenesis. J. Gen. Virol. 5:47-51. von Bonsdorff C-H (1973): The structure of Semliki Forest Virus. Commentationes Scientiarum Fennica. 74:1-53.  Biologicae Societas  WHO Technical Report Series NO. 463 (1971): WI-JO expert committee on biological standardization. 23rd report. World Health Organization, Geneva, Switzerland, p.18. Warner NL and Ovary Z (1970): Biological properties of a mouse IgG2a myeloma protein with anti dinitrophenyl activity. J. Immunol. 105:812-817. Waxiiam MN and Wolinsky 35 (1985a): Detailed immunologic analysis of the structural polypeptides of rubella virus using monoclonal antibodies. Virology 143:153-165. Wanham MN and Wolinsky JS (1985b): A model of the structural organization of rubella virions. Rev. Tnf. Dis. 7:S133-S139. Weil ML, Itabashi HH, Cremer NE, Oshiro LS, Lennette EH and Carnay L (1975): Chronic Progressive Panencephalitis due to rubella simulating SSPE. N. Eng. 3. Med. 292:994-998. Weiss B, Nitchico H, Ghattas I, Wright R and Schlesinger 5 (1989): Evidence for specificity in the encapsidation of Sindbis virus RNAs. 3. Virol. 63:53 10-5318. Weller TH and Neva FA (1962): Propagation in tissue culture of cytopathic agents from patients with a rubellalike illness. Proc. Soc. Exp. Biol. Med. 111:215-225. -125-  Westaway EG, Brinton MA, Gaidamovich SY, Horzinek MC, Igarashi A, Káriáinen L, Lvov DK, Porterfield JS, Russell PK and Trent DW (1985): Flaviviridae. Intervirology 24:183-192. Whiteley PJ and Kapp JA (1989): Tolerance in transgenic mice to a non-MHC self protein is not a result of clonal deletion. In R Meichers (ed): “Progress in Immunology VII”. Berlin: Springer Verlag, pp 826-832. Wolinsky JS (1990): Rubella. In BN Fields and DM Knipe (eds): “Fields Virology” ,2nd ed. New York, NY: Raven Press, pp 8 15-838. Wolinsky JS (1990): Subacute sclerosing panencephalitis, progressive rubella panencephalitis and multifocal leukoencephalopathy. In BH Waksman (ed): “Immunologic Mechanisms in Neurologic and Psychiatric Disease”. New York: Raven Press, pp 259-268. Wolinsky JS, McCarthy M, Allen-Cannady 0, Moore WT, Jin R, Cao 5, Lovett A and Simmons D (1991): Monoclonal antibody-defined epitope map of expressed rubella virus protein domains. J. Virol. 65:3986-3994. Wraith DC (1987): The recognition if influenza A virus infected cells by cytotoxicT lymphcytes. Immunol. Today 8:239-246. Wright KE, Salvato MS and Buchmeier MJ (1989): Neutralizing epitopes of lymphocytic choriomeningitis virus are conformational and require both glycosylation and disulfide bonds for expression. Virology 171:417-426. Yoneda T, Urade M, Sakuda M and Miyazaki T (1986): Altered growth, differentiation and responsiveness to epidermal growth factor of human embryonic mesenchymal cells of palate by persistent rubella virus infection. J. Chin. Invest. 77: 1613-1621. Yoon J-W, Choi D-S, Liang H-C, Baek H-S, Ko I-Y, Jun HS and Gillam S (1991): Induction of an organ specific autoimmune disease, lymphocytic hypophysitis, in hamsters by recombinant rubella virus glycoprotein and prevention of disease by neonatal thymectomy. J. Virol. 66:1210-1214. Young J D-E, Young GPH, Cohn ZA and Lenard J (1983): Interaction of enveloped viruses with bilayer membranes: observations on sendai, influenza, vesicular stomatitis and Semliki Forest viruses. Virology 128:186194. Zhang T, Mauracher CA, Mitchell LA and Tingle AJ (1992): Detection of rubella-specific IgG, 1gM and IgA by immunoblot assays. J. Clin. Microbiol. 30:824-830. Zheng DX, Dickens L, Liu T-Y and Nakhasi ilL (1989): Nucleotide sequence of the 24S subgenomic messenger RNA of a vaccine strain of rubella virus comparison with wild type strain (M33). Gene 82:343-349. -  Zinkemagel RM, Cooper 5, Chambers J, Lazzarin RA, Hengartner 11 and Arnheiter H (1990): Virus-induced autoantibody responses to transgenic viral antigen. Nature 345:68-71. Ziola B, Lund G, Meurman 0 and Salmi A (1983): Circulating immune complexes in patients with acute measles and rubella virus infections. Infect. Immun. 41:578-583.  -126-  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items