Studies of the Epidemiology of Human Herpesvirus 6

Julie Dawn Fox

Submitted for the degree of Doctor of Philosophy University College London ProQuest Number: 10797776

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest

ProQuest 10797776

Published by ProQuest LLC(2018). Copyright of the Dissertation is held by the Author.

ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 Acknowledgements.

I would like to thank Prof. R. S. Tedder and Prof. J. R. Pattison for giving me the opportunity to study for a PhD within the Department of Medical Microbiology and for supervision and helpful discussion during the course of this project. I would also like to thank other members of the Virology Section for help and encouragement.

Thanks are due to Dr M. Contreras (North London Blood Transfusion Centre, London), Dr H. Kangro (St Bartholomew's Hospital, London), Dr H. Holzel (Hospital for Sick Children, London), Dr T. Stammers (The Church Lane Practice, London), Dr M. Hambling (Public Health Laboratory, Leeds) Dr W. Irving (University Hospital, Nottingham) Dr H. Steel (St George's Hospital, London), Dr E. Larson (Clinical Research Centre, London) and Dr P. Venables (The Kennedy Institute of Rheumatology, London) for supplying sera and clinical information. Dr R. Honess (National Institute for Medical Research, London), who sadly died last year, supplied HHV6 control plasmids and gave encouragement and advice on molecular techniques.

2 Abstract

A new human herpesvirus was isolated in 1986 from patients with various lymphoproliferative disorders and was designated human herpesvirus 6 (HHV6). This thesis describes work undertaken to define the epidemiology of HHV6 infection in order to determine any association of HHV6 infection with human disease. The close relationship of HHV6 with human cytomegalovirus and the interaction of HHV6 with other viruses are discussed.

Work on the epidemiology and pathogenic significance of HHV6 depends on having sensitive and specific assays for antibody detection. The development of assays for the detection of HHV6 specific IgM and IgG antibody and the use of these assays to establish antibody prevalence levels in normal and patient population groups are described.

Most adults were found to be HHV6 antibody positive and the majority of HHV6 infections in adults were due to reactivation or reinfection. Children were found to have been infected with HHV6 during the first year of life and the clinical consequences of primary HHV6 infection was investigated. HHV6 antibody prevalence and titre in adult patient groups was compared with controls and the results are discussed in the context of other published serological reports.

In situ hybridization and immunohistochemical staining were used to determine the in vivo tissue tropism of HHV6 and to investigate any potential disease association of HHV6 infection. Amplification of HHV6 DNA by the polymerase chain reaction was used for the detection of HHV6 DNA in saliva samples and peripheral blood mononuclear cells from healthy individuals. A large proportion of salivary gland tissues and saliva samples was found to be HHV6 DNA positive, which may have implications for transmission of the virus.

The results of a prospective study to investigate possible reactivation of HHV6 during pregnancy are presented. Samples were taken from pregnant women at various times during the antenatal and early postnatal period to determine any change in HHV6 antibody profile, HHV6 DNA in peripheral blood samples and shedding of HHV6 in saliva. The postulate is discussed that pregnant women reactivate HHV6 which results in increased salival shedding of the virus during the early postnatal period and thus transmission to their newborn offspring.

3 Contents

Title page ...... 1 Acknowledgements ...... 2 Abstract...... 3 Contents ...... 4 List of Tables ...... 10 List of Figures ...... 13 Layout of thesis ...... 17 Abbreviations ...... 18

Section 1: General Introduction la. TheHerpesviridae family:...... 20-22 la. 1. herpesvirus D N A ...... 20 la. 2. herpes vims proteins ...... 20 la. 3. replication of herpesviruses ...... 21 la. 4. classification of herpesviruses ...... 21 lb. Biological properties of the human herpesviruses: ...... 22-30 lb. 1. herpes simplex vims (HSV) ...... 23 lb. 2. varicella-zoster vims (VZV) ...... 24 lb. 3. Epstein-Barr vims (EBV) ...... 25 lb. 4. human cytomegalovims (CMV) ...... 27 lb. 5. human herpesvirus 6 (HHV6)...... 28 lc. Aims of this study and plan of experimental w ork ...... 31

Section 2: Materials and methods

2a. Tissue culture and HHV6 propagation ...... 32 2b. Indirect immunofluorescence (IF): ...... 34-36 2b. 1. preparation of slides and s e ra ...... 34 2b. 2. detection of anti-HHV6 IgG ...... 36 2b. 3. detection of anti-HHV6 IgM ...... 36 2c. Anti-complementary immunofluorescence (ACIF) for detection of anti-HHV6 IgG ...... 36 2d. Competitive radioimmunoassay (RIA) for detection of anti-HHV6: ...... 37-40 2d. 1. globulin preparation ...... 37 2d. 2. antigen preparation ...... 38

4 2d. 3. preparation of 125I-labelled Ig G ...... 38 2d. 4. detection of anti-HHV6 ...... 39 2e. Complement fixation test for detection of anti-HSV, anti-VZV and anti-CMV: ...... ,40-41 2e. 1. preparation of sensitized sheep red blood cells ...... 40 2e. 2. performance of the test ...... 40 2f. Indirect IF for detection of anti-EBV VC A IgG ...... 41 2g. Solid-phase enzyme immunoassay for detection of anti-CMV IgM ...... 42 2h.In situ hybridization for detection of HHV6 DNA: ...... 42-46 2h. 1. preparation of biotin labelled probes by nick translation ...... 42 2h. 2. reproducibility of the nick translation reaction ...... 43 2h. 3. pre-treatment of cells, slides and tissues ...... 44 2h. 4. hybridization of biotinylated probes to target DNA . . . .45 2h. 5. detection of biotinylated probes ...... 45 2i. Production of monoclonal antibodies against HHV6:...... 46-50 2i. 1. animals and immunisation protocol ...... 46 2i. 2. culture of myeloma cells ...... 47 2i. 3. preparation of feeder cells ...... 47 2i. 4. preparation of spleen cells ...... 47 2i. 5. cell fusion ...... 48 2i. 6. culture of hybrid cells ...... 48 2i. 7. detection of anti-HHV6 secreting hybridomas by indirect I F ...... 48 2i. 8. cloning by limiting dilution ...... 49 2i. 9. culture of hybrid clones ...... 49 2j. Immunohistochemical staining for detection of HHV6- specific antigens: ...... 50-51 2j. 1. pre-treatment of cells, slides and tissues...... 50 2j. 2. the immunohistochemistry reaction ...... 50 2k. The nested polymerase chain reaction (PCR) for detection of HHV6DNA:...... 51-56 2k. 1. design and preparation of oligonucleotide primers .... 51 2k. 2. preparation of DNA from tissue, mononuclear cell and oropharyngeal samples ...... 54 2k. 3. first round PCR reaction ...... 55 2k. 4. second round PCR reaction...... 55 2k. 5. analysis of PCR products ...... 56 21. Statistical methods ...... 56

5 Section 3: Results

3a. Serological assays for the detection of anti-HHV6

3a. 1. Introduction ...... 57 3a. 2. Detection of anti-HHV6 IgG by IF : ...... 59-64 3a. 2. i. titres and scoring system for indirect I F ...... 59 3a. 2. ii. reproducibility of indirect I F ...... 62 3a. 2. iii. comparison of indirect IF and ACIF for detection of anti-HHV6 IgG ...... 62 3a. 3. Detection of anti-HHV6 by competitive RIA: ...... 65-69 3a. 3. i. optimization of the assay ...... 65 3a. 3. ii. titration ...... 65 3a. 3. iii. reproducibility ...... 69 3a. 4. Anti-HHV6 IgG in serum samples from blood donors: ...... 69-79 3a. 4. i. by indirect I F ...... 69 3a. 4. ii. by competitive R IA ...... 69 3a. 4. iii. correlation between indirect IF and competitive R IA ...... 75 3a. 5. Anti-HHV6 IgG in serum samples from children ...... 79 3a. 6. Comparison of HHV6 antibody and antibody to the other human herpesviruses in sera from adults and children ...... 79 3a. 7. Detection of anti-HHV6 IgM by indirect IF : ...... 86-90 3a. 7. i. titres and scoring system ...... 86 3a. 7. ii. specificity and reproducibility ...... 86 3a. 8. A study of HHV6 antibody in sera from pregnant women and female blood donors ...... 90 3a. 9. Discussion ...... 92

3b. HHV6 antibody in patient groups

3b. 1. A study of HHV6 antibody in children with roseola infantum ...... 94 3b. 2. Reactivation of HHV6 in 3 patients ...... 99 3b. 3. HHV6 and chronic fatigue syndrome: 102 - 106 3b. 3. i. anti-HHV6 IgG in a group of Paul-Bunnell negative glandular fever patients ...... 103 3b. 3. ii. anti-HHV6 in a group of chronic fatigue syndrome patients ...... 106 3b. 4. HHV6 in AIDS patients: 106 - 116

6 3b. 4. i. anti-HHV6 IgG in sera from HIV-1 infected and uninfected homosexual m en ...... 110 3b. 4. ii. anti-HHV6 in sera from HIV-1 infected and uninfected haemophiliacs ...... 112 3b. 5. Discussion ...... 116

Section 3c. Detection of HHV6 DNA and protein in healthy individuals

3c. 1. Nested PCR for detection of HHV6 DNA: ...... 119-123 3c. 1. i. selection of primer sequences ...... 120 3c. 1. ii. optimization of reaction conditions ...... 120 3c. 2. In situ hybridization for detection of HHV6 D N A : ...... 123-127 3c. 2. i. reproducibility of the biotinylation reaction ...... 125 3c. 2. ii. detection of the biotinylated probe ...... 125 3c. 3. Immunohistochemical staining for detection of HHV6 protein: . . 122-128 3c. 3. i. hybridomas secreting anti-HHV6 ...... ,127 3c. 3. ii. detection of the primary monoclonal antibody ...... 128 3c. 4. Detection of HHV6 DNA and protein in salivary gland tissues: ...... 128-132 3c. 4. i. HHV6 DNA by nested PC R ...... 128 3c. 4. ii. HHV6 DNA by in situ hybridization ...... 130 3c. 4. iii. HHV6 protein by immunohistochemical staining ...... 130 3c. 5. Detection of HHV6 DNA by nested PCR in peripheral blood mononuclear cells and oropharyngeal samples ...... 133 3c. 6. Discussion ...... 134

3d. Detection of HHV6-specific DNA and protein in samples from patient groups

3d. 1. HHV6 DNA by in situ hybridization in postmortem tissues from cases of sudden infant death syndrom e ...... 136 3d. 2. HHV6 DNA by nested PCR in tissues from patients with inflammatory bowel disease ...... 137 3d. 3. HHV6 DNA by nested PCR in gut lymphomas ...... 145 3d. 4. HHV6 DNA by nested PCR in malignant tumours derived from nervous tissue ...... 145 3d. 5. HHV6 protein by immunohistochemistry in tissues from Sjogren's syndrome patients ...... 145

7 3d. 6. Discussion 146

3e. A longitudinal study of HHV6 in pregnancy

3e. 1. Introduction and aims of this study ...... 150 3e. 2. Sample collection and analysis of data ...... 151 3e. 3. HHV6 antibody in sera from a group of 70 pregnant wom en ...... 151 3e. 4. HHV6 antibody in sera from 57 women followed through pregnancy ...... 154 3e. 5. HHV6 DNA levels in peripheral blood mononuclear cells and throatwashings from a group of 70 pregnant women ...... 156 3e. 6. HHV6 DNA in peripheral blood mononuclear cells and throatwashings from 57 women followed through pregnancy 159 3e. 7. A summary of results for HHV6 in pregnancy ...... 161 3e. 8. Discussion ...... 161

Section 4: General Discussion

4a. Isolation of HH V 6 ...... 164-175 4a. 1. ultrastructural characterization ...... 168 4a. 2. cell tropism and in vitro growth characteristics...... 171 4a. 3. sequence variation amongst HHV6 isolates ...... 174 4a. 4. antigenic differences between HHV6 isolates ...... 174 4b. Detection of anti-HHV6 in healthy individuals ...... 175 4c. Tissue localization of HH V 6 ...... 178 4d. HHV6 in pregnancy - implications for transmission ...... 183 4e. HHV6 potential disease associations: ...... 184 - 197 4e. 1. roseola infantum ...... 184 4e. 2. sudden infant death syndrome ...... 187 4e. 3. chronic fatigue syndrome ...... 187 4e. 4. hepatitis...... 189 4e. 5. acquired immune deficiency syndrome ...... 189 4e. 6. disease and graft rejection in transplant recipients ...... 192 4e. 7. benign lymphoproliferative disorders ...... 194 4e. 8. tumours derived from lymphoid and nervous tissue ...... 195

8 4e. 9. inflammatory bowel disease ...... 196 4f. Interaction of HHV6 with the other human herpesviruses...... 197 4g. Natural history and pathogenesis of HHV6 infection ...... 199 4h. Prevention and treatment of HHV6 infection ...... 200 4i. HHV6 molecular studies ...... 202 4j. Classification of HHV6 ...... 203 4k. Conclusions ...... 205 Appendix A: buffer and media recipes ...... 206 Appendix B: publications arising from this thesis ...... 211 References ...... 213

9 List of Tables

Table 2.1 Primer sequences, anneal temperatures and expected product sizes for herpesvirus DNAs by nested PC R ...... 53

Table 3a. 1 Range of titres for each group of sera scored for anti-HHV6 IgG by indirect IF ...... 61

Table 3a. 2 Radiolabel bound to HHV6 antigen in the presence of different animal sera .... 66

Table 3a. 3 Anti-HHV6 IgG by indirect IF in sera from male and female blood donors ...... 72

Table 3a. 4 Scoring system for anti-HHV6 by competitive RIA ...... 74

Table 3a. 5 Anti-HHV6 by competitive RIA in sera from male and female blood donors ...... 76

Table 3a. 6 Anti-HHV6 IgG by indirect IF in sera from 252 children aged 1-17 years ...... 80

Table 3a. 7 Anti-HHV6 IgG by indirect IF in sera from 112 children under 1 year of ag e ...... 81

Table 3a. 8 Range of titres for each group of sera scored for anti-HHV6 IgM by I F ...... 88

Table 3a. 9 Anti-HHV6 by indirect IF and RF by latex-agglutination before and after treatment with RF-absorbans ...... 89

10 Table 3a. 10 Antibody to HHV6 in sera from pregnant women and female blood donors ...... 91

Table 3b. 1 Antibody to HHV6 and CMV in sera from 11 children with a clinical diagnosis of roseola infantum ...... 95

Table 3b. 2 Antibody to HHV6 and CMV in sera from 7 children with pyrexia and rash...... 96

Table 3b. 3 A comparison of anti-HHV6 by indirect IF and competitive RIA in sera from children with pyrexia and rash ...... 97

Table 3b. 4 Antibody to HHV6 in 3 cases of HHV6 reactivation ...... 101

Table 3b. 5 Anti-HHV6 IgG by indirect IF in sera from Paul-Bunnell negative glandular fever patients and controls ...... 105

Table 3b. 6 Anti-HHV6 by indirect IF in sera from chronic fatigue syndrome patients and controls ...... 107

Table 3b. 7 Anti-HHV6 by competitive RIA in sera from chronic fatigue syndrome patients and controls ...... 108

Table 3b. 8 Anti-HHV6 IgG by indirect IF in sera from HIV-1 infected and uninfected homosexual m en ...... I l l

Table 3b. 9 Anti-HHV6 IgG by indirect IF in sera from HIV-1 infected and uninfected haemophiliacs ...... 113

11 Table 3b. 10 Anti-HHV6 by competitive RIA in sera from HIV-1 infected and uninfected haemophiliacs ...... 114

Table 3d. 1 Detection of HHV6, CMV and EBV DNA by nested PCR in tissues from inflammatory bowel disease patients and controls ...... 143

Table 3e. 1 HHV6 antibody in serum samples from a group of 70 pregnant women ...... 152

Table 3e. 2 Statistical analyses of HHV6 antibody prevalence in sera from women taken at various times during pregnancy and the postnatal period ...... 155

Table 3e. 3 HHV6 DNA by nested PCR in peripheral blood mononuclear cells and throatwashings from a group of 70 pregnant women ...... 157

Table 3e. 4 Statistical analyses of HHV6 DNA prevalence in peripheral blood mononuclear cells and throatwashings taken at various times during pregnancy and the postnatal period ...... 160

Table 4.1 HHV6 isolation from adults ...... 165-167

Table 4. 2 HHV6 isolation from children ...... 169-170

12 List of Figures

Figure 1.1 Enveloped virus particles in cytoplasmic vesicles ...... 29

Figure 3a. 1 HHV6 CPE in the J Jhan T-cell line...... 60

Figure 3a. 2 Section through a pellet of HHV6 infected J Jhan cells ...... 60

Figure 3a. 3 A comparison of anti-HHV6 IgG score and titre by indirect IF for 25 blood donor sera ...... 61

Figure 3a. 4 Photomicrographs of indirect IF screening of sera for anti-HHV6 IgG ...... 63

Figure 3a. 5 Photomicrographs of the fluorescence produced by an anti-HHV6 IgG positive serum by ACIF and indirect I F ...... 64

Figure 3a. 6 Titration of 5 categories of sera in the anti-HHV6 competitive R IA ...... 67

Figure 3a. 7 Titration of 5 sera in the anti-HHV6 competitive RIA: assessment of variation between duplicate reactions in one assay ...... 68

Figure 3a. 8 Results for 5 control sera screened for anti-HHV6 by competitive RIA: assessment of inter-test variability ...... 70

Figure 3a. 9 Anti-HHV6 IgG by indirect IF in sera from 96 blood donors ...... 71

Figure 3a. 10 Anti-HHV6 IgG score by IF compared with age for 93 blood donor sera ...... 73

13 Figure 3a. 11 Anti-HHV6 by competitive RIA in sera from 96 blood donors ...... 74

Figure 3a. 12 Anti-HHV6 score by competitive RIA compared with age for 93 blood donor sera ...... 77

Figure 3a. 13 Correlation between indirect IF and competitive RIA for anti-HHV6 in sera from 96 blood donors ...... 78

Figure 3a. 14 Comparison of score and titre for anti-EBY VCA IgG by indirect IF in 37 blood donor sera ...... 82

Figure 3a. 15 Comparison of anti-HHV6 IgG by indirect IF and antibody to the other human herpesviruses in 96 blood donor sera ...... 83

Figure 3a. 16 Comparison of anti-HHV6 by competitive RIA and antibody to the other human herpesviruses in 96 blood donor sera ...... 84

Figure 3a. 17 Comparison of anti-HHV6 IgG by IF and antibody to the other human herpesviruses in sera from 28 children ...... 85

Figure 3a. 18 Photomicrographs of indirect IF screening of sera for anti-HHV6 IgM ...... 87

Figure 3a. 19 A comparison of anti-HHV6 IgM score and titre by indirect IF for 15 sera ...... 88

Figure 3b. 1 % inhibition in the anti-HHV6 competitive RIA for sera from chronic fatigue patients and controls ...... 109

14 Figure 3b. 2 % inhibition in the anti-HHV6 competitive RIA for sera from HIV-1 infected and uninfected haemophiliacs ...... 115

Figure 3c. 1 The effect of primer and magnesium concentration on the amount of amplified PCR product ...... 122

Figure 3c. 2 Nested PCR amplification of HHV6 plasmid using optimum reaction conditions ...... 124

Figure 3c. 3 Titration and detection of biotinylated probes on a nitrocellulose filter ...... 126

Figure 3c. 4 Photomicrographs of the fluorescence produced by 3 HHV6 monoclonal antibodies by indirect IF ...... 129

Figure 3c. 5 HHV6 DNA-specific staining by in situ hybridization in submandibular and parotid salivary glands ...... 131

Figure 3c. 6 HHV6-specific protein staining with monoclonal antibodies in submandibular and parotid salivary g lan d s ...... 132

Figure 3d. 1 CMV DNA-specific staining by in situ hybridization in lung tissue from an AIDS patient ...... 138

Figure 3d. 2 HHV6 infected tissue culture cells probed for virus-specific DNA by in situ hybridization ...... 138

Figure 3d. 3 Single round PCR amplification of HHV6 and B-globin sequences in DNA extracted from intestinal tissue ...... 141

15 Figure 3d. 4 Nested PCR amplification of HHV6 DNA in tissue samples from inflammatory bowel disease patients ...... 142

Figure 3d. 5 Parotid gland lymphoproliferative lesions screened for HHV6 and CMV-specific proteins using monoclonal antibodies ...... 147

Figure 3e. 1 HHV6 antibody in sera from a group of 70 pregnant women ...... 153

Figure 3e. 2 HHV6 DNA by nested PCR in peripheral blood mononuclear cells and throatwashings from a group of 70 pregnant women ...... 158

16 Layout of thesis

This thesis has been divided into four main sections.

Section 1: General Introduction.

This includes general information about the Herpesviridae family, epidemiology of the human herpesviruses HSV, CMV, VZV and EBV and background to the identification and isolation of HHV6.

Section 2: Materials and Methods

Section 3: Results

The experimental results have been grouped into five sub-sections (3a-3e respectively) each of which covers a particular area of study. Where appropriate, a short introduction covering the background work to each area of study and a discussion of experimental results in the context of other related published work has been included in each sub section.

Section 4: General Discussion

This section includes a comparison of the experimental results obtained for HHV6 with that for the other human herpesviruses, conclusions to be drawn from the experimental work and a summary of published work on HHV6. Classification of HHV6 based on all available molecular and epidemiological data is discussed.

17 Abbreviations used in this thesis

A = adenine AIDS = acquired immune deficiency syndrome p2M = p2 Microglobulin BCDP = 5-bromo-4-chloro-3-indoyl phosphate (disodium salt) BSA = bovine serum albumin bp = base pairs C = cytosine CD = Crohns disease CD4 = cluster differentiation 4 CFS = chronic fatigue syndrome CF(T) = complement fixation (test) CMI = cell mediated immunity CMV = human cytomegalovirus CNS = central nervous system CPE = cytopathic effect CSF = cerebrospinal fluid DAB = diaminobenzidine tetrahydrochloride DMF = N,N-dimethyl formamide DMSO = dimethyl sulphoxide DNA = deoxyribonucleic acid DNase = deoxyribonuclease EA = early antigen EBV = Epstein-Barr virus EDTA = disodium ethylenediamine tetra-acetate. 2H20 ds = double stranded EIA = enzyme immunoassay ELISA = enzyme-linked immunosorbent assay EM = electron microscopy FPLC = fast performance liquid chromatography Fc = crystallizable fraction of Ig G = guanine FCS = foetal calf serum G+C = % molar content of guanine and cytosine HBLV = human B-lymphotropic virus HEPES = N-2-hydroxyethylpiperazine-N'-2-ethanesulphonic acid HIV = human immunodeficiency virus HLA = human leukocyte antigen

18 HSV = herpes simplex virus HTLV-1 = human T-lymphotropic virus type 1 HVS = herpesvirus simairi IBD = inflammatory bowel disease IF = immunofluorescence Ig = immunoglobulin IL-2 = interleukin 2 IM = infectious mononucleosis i.p. = intraperitoneal i.v. = intravenous Kbp = kilobase pair KDa = kilodalton MAb = monoclonal antibody MCP = major capsid protein ME = myalgic encephalomyelitis Mr = relative molecular mass mRNA = messenger RNA NBT = nitroblue tetrazolium ND = not done OD = optical density PBMC = peripheral blood mononuclear cell(s) PBNGF = Paul-Bunnell negative glandular fever PBS = phosphate buffered saline PHA = phytohaemagglutinin PVFS = post viral fatigue syndrome RIA = radioimmunoassay RIPA = radioimmunoprecipitation of antigen RNA = ribonucleic acid RNase = ribonuclease RPMI = Roswell Park Memorial Institute sd = standard deviation SDS = sodium dodecyl sulphate ss = single stranded SS = Sjogrens syndrome T = thymine Tris = Tris (hydroxymethyl) methylamine TW = throatwashings UC = ulcerative colitis VCA = viral capsid antigen

19 Section 1: General Introduction

Herpesviruses are ubiquitous. Over one hundred different herpesviruses have been identified in a wide range of animal species and those examined to date are capable of latency in their natural hosts. In cells harbouring latent virus only a small subset of the viral genes are transcribed and the virus is subsequently capable of reactivation from this latent state. Human herpesvirus 6 is one of the most recently identified herpesviruses and is the subject of this thesis. la. The Herpesviridae family :

Membership in the family Herpesviridae is based on the architecture of the virion (Mj. 100 x 109, buoyant density in caesium chloride 1.20-1.29 g / ml) and all herpesviruses have similar morphology by electron microscopy (EM). The herpesvirus particle is spherical, enveloped and 120-200 nm in diameter. The particle is made up of a core-protein spool around which the DNA is wrapped, a capsid (100-110 nm diameter) comprising 162 hollow prismatic capsomeres arranged in icosahedral symmetry, an amorphous, globular tegument surrounding the capsid and a pleomorphic envelope (lipid bilayer) external to the tegument in which are embedded the viral glycoproteins. Herpesviruses are sensitive to ether and lipid solvents, extremes of pH and are relatively thermolabile (Roizman et al ., 1981). la. 1. herpesvirus DNA

Herpesviruses have, by definition, one molecule of linear double-stranded DNA (Mj. 70-150 x 106, 120-240 Kbp, average G+C 32-75%). The genomes from different herpesviruses have single or multiple direct terminal repeats which vary in length from many kilobases to fewer than 50 bases. In many instances sequences from one or both termini are reiterated internally in inverted orientation and six groups of herpesviruses (A-F) have been described based on the presence and location of these sequences (Roizman, 1990). Since the terminal sequences contain signals for cleaving and packaging of concatameric DNA, the presence of these internal copies results in alternative sites for initiation of DNA packaging. Together with intermolecular recombination this provides a mechanism whereby different virus particles in a single population contain different isomeric forms of the genome. la. 2. herpesvirus proteins

All herpesviruses studied so far specify a large number of enzymes involved in nucleic acid metabolism, DNA synthesis and perhaps also protein processing. Herpesviruses

20 code for > 20 structural proteins (Mr 12-220 x 103) and > 6 glycoproteins, most of the mRNA transcripts are not spliced. Some proteins are conserved amongst the herpesviruses (examples include the major capsid antigen, DNA polymerase, ribonucleotide reductase and glycoprotein B). The glycoproteins seem to be the target for antibody-mediated neutralization and are responsible for entry functions of the virus and modifications of the infected cell surface. la. 3. replication of herpesviruses

Little is known about the cellular receptors and viral attachment proteins for herpesviruses. Fusion of the viral membrane with the cellular membrane is thought to take place at the cell surface rather than in a cytoplasmic vesicle following endocytosis since lysosomal inhibitors do not inhibit viral penetration. After penetration, the nucleocapsid and associated tegument proteins are transported to the nucleus where components of the virion are involved in early shut-off of host macromolecular synthesis. Herpesvirus DNAs circularize immediately upon release from capsids into the nuclei of infected cells and immediate early (a) genes are transcribed by host RNA polymerase II, most of the a gene products have been confirmed as having regulatory functions. The initial transcription is independent of viral protein synthesis and is not absolutely dependent on any virion component. Delayed early (p) gene transcripts encode for enzymes involved in DNA replication and nucleotide metabolism. DNA synthesis occurs at one or more origins of replication by the Tolling circle' method producing concatemeric DNA. The late (y) genes are transcribed after DNA replication and encode many of the virion structural proteins. The concatemeric DNA is cleaved and packaged into pre-formed capsid structures. It is now accepted that the virus buds through the nuclear membrane and enveloped particles can be found in the peri-nuclear space. Release of the viral particle is thought to occur by reverse endocytosis. la. 4. classification of herpesviruses

Historically the herpesviruses have been classified into 3 sub-families, the Alphaherpesvirinae, Betaherpesvirinae and Gammaherpesvirinae based on their biological properties. The validity of this classification has been reinforced, in most cases, by nucleic acid hybridization and analysis of nucleotide sequence data (Minson, 1989).

Alphaherpesviruses have a short replication cycle (usually hours rather than days) and spread rapidly in tissue culture causing cell lysis. Although species-specific in vivo they will productively infect fibroblasts derived from many species. In the natural host they usually cause only mild illness of short duration. Infection notably involves the epithelium of skin and mucosal surfaces although most evidence suggests that

21 alphaherpesviruses are capable of latent infection of nerve tissue. Herpes simplex is the most studied alphaherpesvirus and may be considered as the prototype virus for this sub-family.

Betaherpesviruses have a long replication cycle and thus a slow spread of infection in tissue culture. The cytopathic effect (CPE) is typified by giant cell formation. They tend to be species-specific in both in vivo and in vitro infections. The in vivo infection is systemic, of considerable duration and virus can be isolated from multiple secretions over a period of months or years. They have larger genomes than alpha or gammaherpesviruses (typically 180-200 Kbp). Betaherpesvirus, like the alpha and gammaherpesviruses, persist and frequently reactivate although the biologically relevant site of persistence is unclear. Human cytomegalovirus is the prototype virus for this sub-family.

Gammaherpesviruses are termed the 'lymphotropic herpesviruses" and are associated with lymphoproliferative diseases in their natural or experimental hosts. Lymphoid cell lines containing multiple copies of the viral DNA can frequently be established from peripheral blood monocytes in the acute or convalescent stage of infection. There are both B-cell lymphotropic (eg. Epstein-Barr virus) and T-cell lymphotropic (eg. herpesvirus saimiri) gammaherpesviruses but these tropisms are not mutually exclusive. Lymphocytes are generally non-permissive or semi-permissive for these viruses and many of them will productively infect fibroblasts. It has been suggested that lymphocytes or their progenitor stem cells are the site of latency of gammaherpesviruses. lb. Biological properties of the human herpesviruses:

The ability of herpesviruses to become latent in their natural host and subsequently reactivate has led to much research interest in the field of human herpesvirus epidemiology. The latent virus forms a reservoir in the immune host and subsequent reactivation ensures that herpesviruses are not dependent on a large number of susceptible individuals in a population in order to persist. Research has particularly centred on the mechanism of herpesvirus latency and the clinical outcome of primary infection, reactivation and reinfection in children and adults in the normal population and in patient groups.

Until 1986 there were 5 recognised human herpesviruses, herpes simplex virus 1 (HSV 1 or human herpesvirus 1), herpes simplex virus 2 (HSV 2 or human herpesvirus 2), varicella-zoster virus (VZV or human herpesvirus 3), Epstein-Barr virus (EBV or human herpesvirus 4) and human cytomegalovirus (CMV or human herpesvirus 5). In 1986 a new herpesvirus (later designated human herpesvirus 6) was identified

22 (Salahuddin et al ., 1986). Studies of the epidemiology of human herpesvirus 6 are the subject of this thesis. The isolation of a seventh human herpesvirus has also been reported (Frenkel et al. 1990) but, as yet, little is known about this virus. The relationship between human herpesvirus 6 and this newly isolated seventh human herpesvirus is discussed in Section 4 (4j.). lb. 1. herpes simplex virus

Herpes simplex virus is an alphaherpesvirus which is seen as two serotypes (HSV 1 and HSV 2) based on minor antigenic differences. The virion polypeptides and viral encoded proteins have been extensively studied (Knipe, 1989, Roizman and Sears, 1990) and two glycoproteins which were first identified in HSV (gB and gH) have homologues in members of all subfamilies of herpesviruses. The HSV type specific glycoproteins gC and gE are thought to be involved in attachment to cellular receptors and glycoproteins gl and gE have been implicated in the production of Fc receptors on the surface of HSV infected cells (evidence for this is reviewed by Roizman and Sears, 1990, Whitley, 1990). Both HSV 1 and HSV 2 are rapidly growing cytolytic viruses which invariably destroy the cells which they infect and in which they multiply. They are, however capable of latent infection of neuronal tissue and have been studied as models for herpesvirus latency (discussed in Roizman and Sears, 1987, Roizman and Sears, 1990 and Whitley, 1990). Reactivation and recrudescence of HSV has been associated with fever, menstruation, irradiation, secondary infections and other cofactors and is a common sequel to immunosuppressive therapy. The frequency and severity of HSV lesions is increased in individuals infected with human immunodeficiency virus (HIV).

Infection with HSV 1 typically occurs during the first six years of life as a consequence of exposure to an individual secreting the virus, probably in saliva. The clinical outcome in these children with primary infection is usually a mild stomatitis which may be unrecognised. Where primary HSV 1 infection is delayed until adolescence a common finding is stomato-gingivitis and pharyngitis in association with a mononucleosis syndrome. Recurrent HSV 1 infection is common and usually manifests clinically as lesions at the mucocutaneous junction of the lip. In rare cases primary or reactivated HSV 1 causes serious and sometimes lethal infections. In particular, primary infection of neonates may result in acute hepatitis or herpetic meningo-encephalitis. In adults rare clinical manifestations of HSV 1 primary infection or reactivation include herpetic keratoconjunctivitis and encephalitis.

In the case of HSV 2, infection usually occurs at birth or in late adolescence during the early years of sexual activity. The virus is transmitted by sexual contact and lesions

23 associated with primary infection or reactivation most commonly occur in the genital area. Clinical symptoms associated with HSV infection are frequently noted during pregnancy (Whitley, 1990) but transplacental infection of a foetus by HSV is probably rare. Neonatal herpes infection resulting from acquisition of HSV 2 in the birth canal occasionally causes a generalized infection involving the viscera and or central nervous system (CNS). These cases occur rarely after reactivation but occur much more commonly after primary infection in the mother. The mortality in such cases is high (Stagno and Whitley, 1985a, Whitley, 1990a). It has been proposed that HSV 2 may be a cofactor in the development of human papilloma virus associated cervical neoplasias (zur Hausen, 1982). A more complete discussion of the clinical manifestations of HSV infection is given in Minson (1989), Longson (1990). lb. 2. varicella-zoster virus

A comparison of the DNA sequence of HSV 1 and VZV revealed similarity in the gene layout of the two viruses suggesting a common ancestry (Davison and Scott, 1986, McGoech et al., 1988). This confirms the biological classification of the two viruses in the same sub-family (Alphaherpesvirinae). Varicella-zoster virus has similar growth properties to HSV, although it grows more slowly in vitro, and is thought to produce latent infection of dorsal root ganglia (Gilden, 1983).

Primary VZV infection causes chickenpox (varicella) a common exanthematous disease of childhood, the majority of infections occurring in children aged 4-10 years. Most individuals in temperate climates contract varicella during childhood and an airborne route of infection for VZV would seem to be most likely. The symptoms of fever, malaise and scalp and trunk lesions become more severe when primary infection is delayed until later in life. The most frequent complication of primary VZV infection in the normal host is due to secondary bacterial infections but in adults and immunocompromised individuals potentially fatal varicella pneumonia, encephalitis and other severe complications of primary infection are more common (Feldman et al., 1975). It is thought that approximately 10% of young adults are still susceptible to varicella (Heath and Kangro, 1990) which has important implications for health care workers and pregnant women. In particular, pregnant women are at increased risk from VZV pneumonitis and the virus may be transmitted to the foetus in utero. Congenital malformation of the foetus has also been reported when the mother suffers a primary infection during the first trimester (Stagno and Whitley, 1985a, Gershon, 1990). Neonatal varicella infection may occur as a consequence of primary infection in the mother less than 8 days before delivery (Heath and Kangro, 1990). A VZV infected neonate may suffer from pneumonia and visceral disease and the fatality rate in such cases is high.

24 Shingles (herpes zoster) is a common manifestation of VZV recurrence (Straus et al., 1984). Recurrent VZV infection usually occurs in older or immunocompromised individuals and the symptoms may persist for months or even years. The most common complication of herpes zoster is post-herpetic neuralgia which may occur in up to 50 % of elderly patients, the prognosis, however is generally good in immunocompetent individuals. Herpes zoster carries no increased risk to a pregnant woman and VZV is only rarely transmitted to the foetus during recurrent infection (Gershon, 1990). Immunosuppressed patients may suffer from severe disseminated zoster with or without visceral involvement. Although the duration of symptoms is often prolonged recurrence of VZV does not usually prove fatal in these individuals (Heath and Kangro, 1990, Gelb, 1990). lb. 3. Epstein-Barr virus

Epstein-Barr virus is a B-lymphotropic gammaherpesvirus which has been fully sequenced (Baer et al ., 1984) and is the only human herpesvirus for which the viral attachment protein and cellular receptor is known. The virus attaches to the C3d complement receptor, CR2, of cells (Fingeroth et al ., 1984) via the 350 / 220 glycoprotein complex. The in vitro host range of EBV is restricted to a few human and primate cell lines. Epstein-Barr virus will infect and lyse a small proportion of the human epithelial cells expressing CR2 in tissue culture, resulting in the production of infectious progeny and death of the cell. Following infection of B-lymphocytes in vitro the virus establishes a latent type infection with immortalization of the cell and a restricted viral gene expression (reviewed in Miller, 1990). The viral DNA persists in episomal form in these latently infected cells.

There are two peaks of seroconversion for EBV in the UK, the first occurs at 1-6 years and the second at 14-20 years. Viral transmission most commonly occurs by the oral route although EBV may also be acquired by sexual contact and blood transfusion. In developing countries most individuals are infected with EBV before 2 years of age and worldwide, 80-90 % of adults have antibody to the virus (Crawford and Edwards, 1990). After seroconversion the majority of individuals shed EBV in saliva continuously or sporadically. The source of this shed virus is lytically infected epithelial cells in the pharynx (Sixbey et al., 1984, Young et al., 1986, Allday and Crawford, 1988). In infected individuals a small number of peripheral blood B-cells carry the EBV genome in its 'latent" state with only limited viral expression .

Most primary EBV infections are subclinical. If delayed until adolescence primary EBV infection results in acute infectious mononucleosis (IM) in 25-50 % of individuals (Henle and Henle 1966, Niederman et al., 1970). Infectious mononucleosis (also known

25 as glandular fever) is characterized by lymphadenopathy, lymphocytosis and 'atypical' mononuclear cells in the peripheral blood. Autoantibodies are a hallmark of IM and this is exploited in the Paul-Bunnell test for heterophile antibody which is used for laboratory confirmation of the disease. The clinical symptoms associated with IM include sore throat, headache and fever and the associated fatigue and malaise may persist for long periods of time. Complications of IM such as overt jaundice, myalgia and neurological symptoms are rare and usually only occur when infection has been delayed until the third decade or later. There have been reports of fatal lymphoproliferative disease in pregnant women with IM but effects on the foetus when the mother seroconverts for EBV during pregnancy are rare (Fleisher and Bolognese, 1983, Stagno and Whitley, 1985). Infected individuals may reactivate EBV subclinically or have a relapse of their IM symptoms (especially when immunosuppressed) with occasional morbid complications. Individuals with X-linked lymphoproliferative syndrome fail to mount a cell-mediated immune response to EBV and account for more than half of the fatal IM cases reported (Crawford and Edwards, 1990). There have also been reports of cases of post viral fatigue syndrome associated with a reactivation of EBV (Straus et al., 1985, Hotchin et al., 1989).

Epstein-Barr virus has been implicated as a cofactor in autoimmune diseases such as rheumatoid arthritis and Sjogren's syndrome (SS, reviewed by Fox, 1988) and a high incidence of EBV DNA sequences in tissue specimens from Hodgkin's disease has also been reported (Herbst et al., 1990). The virus has been aetiologically associated with two well defined cancers in man, endemic Burkitt's lymphoma (BL) and undifferentiated nasopharyngeal carcinoma (NPC). Virus-specific DNA and nuclear antigens have been identified in these tumours (zur Hausen et al., 1970, Wolf et al., 1973) but the populations at risk are almost 100 % EBV infected and, as it appears that other cofactors are involved, a causative role for EBV has been difficult to prove. Primary EBV infection of the immunosuppressed may result in atypical IM with gastrointestinal symptoms (Crawford and Edwards, 1990) and renal transplant patients additionally may show graft rejection. There is an increased incidence of EBV positive B-cell lymphoproliferative disorders and neoplasias in graft recipients receiving T-cell suppressive treatments (Crawford et al., 1980, Cleary et al., 1984) and EBV associated oral hairy leukoplakia (Greenspan et al., 1985) and pneumonitis in acquired immune deficiency syndrome (AIDS) patients have also been reported. A review of EBV associated lymphoproliferative diseases and tumours is given in Purtilo et al. (1985), Epstein (1986), Crawford and Edwards (1990), Miller (1990).

26 lb. 4. human cytomegalovirus

Human CMV is a betaherpesvirus with a highly complex genome (230 Kbp), some areas of which are homologous with regions of the human DNA sequence (Rliger at al, 1984, Chee et al., 1990). The virus is associated with B2 microglobulin (B2M) in body fluids and because of this it has been suggested that the class I HLA molecule may be the cell surface receptor for CMV as well as B2M (McKeating et a l, 1986, Grundy et al., 1987). Human fibroblasts are the only cells fully permissive for CMV in vitro and several days after infection CMV induces IgG isotype-specific Fc receptors in these infected fibroblasts (Keller et al., 1976). The in vivo tissue tropism for CMV is much wider. It has been proposed that CMV latency is associated with T- and or B- lymphocytes which allow only limited viral expression (discussed in Braun and Reiser, 1986, Stinski 1990). 'Latent" CMV has been reported in arterial walls of atherosclerosis patients (Hendrix et al., 1990).

In countries with good social circumstances approximately 40 % of adolescents have been infected with CMV (Griffiths and Baboonian, 1984, Griffiths, 1990), infection commonly occurring at or soon after birth (Stagno et al., 1980). Primary CMV infection is usually asymptomatic but the clinical outcome may be more severe in older individuals when sexual contact is thought to be an important mode of transmission of the virus (Handsfield et al., 1985). Cytomegalovirus can be transmitted by blood transfusion and a mononucleosis-like illness may be a result of primary infection acquired in this way (Jordan et al., 1973, Adler, 1983). Viral shedding into saliva and genital secretions may persist for many years.

Cytomegalovirus is one of the most common viruses known to be transmitted in utero (Ahlfors et al., 1984, Stagno and Whitley, 1985, Stagno, 1990) and in pregnant women reactivation may result in renewed shedding of infectious virus (Stagno et al., 1975, Pass et al, 1982, Pass, 1985). Congenital CMV infection may occur as a result of primary infection or reactivation in the mother but primary infection has most often been associated with serious clinical disease and birth defects in the infant (Stagno et al., 1982, Preece et al., 1984). The clinical symptoms of newborns infected in utero include growth retardation and encephalitis and infected babies who are asymptomatic at birth may suffer from hearing loss and impaired intellect later in life. Cytomegalovirus may also be acquired perinatally or neonatally and genital secretions, breast milk and saliva are thought to be the most common sources of infection (Stagno et a l, 1980, Dworsky et al., 1983); most of these infections are asymptomatic although primary CMV infection is an important cause of pneumonia in the very young. Immunocompromised individuals may suffer severe clinical manifestations of CMV primary infection or reactivation such as pneumonia (discussed in Griffiths, 1990,

27 Stinski, 1990). Primary CMV infection, reinfection and reactivation have been implicated in renal transplant rejection (Glenn, 1981, Grundy et al., 1988). Kaposi's sarcoma in AIDS patients has been associated serologically and virologically with CMV infection (discussed in Giraldo et al ., 1989). lb. 5. human herpesvirus 6

In 1986, Salahuddin and colleagues reported isolation of a new human herpesvirus from six patients with lymphoproliferative disorders, two of these patients were HIV infected with AIDS related symptoms (Salahuddin et al ., 1986). The HIV-uninfected individuals suffered from angioimmunoblastic lymphadenopathy, cutaneous T-cell lymphoma and immunoblastic lymphoma, respectively. Ultrastructural characterization of the new isolates confirmed their classification as members of the Herpesviridae family (Biberfeld et al., 1987). The six isolates were antigenically distinct from other human herpesviruses but closely related to each other and were reported to infect B- lymphocytes selectively in tissue culture, they were thus designated human B- lymphotropic virus (HBLV). The novel nature of this herpesvirus was confirmed by Southern blotting, a 9 Kbp portion of cloned HBLV DNA did not hybridize with DNA from the other human herpesviruses and no hybridization of HSV-, VZV-, CMV-, EBV- and herpesvirus saimiri (HVS)-specific probes with HBLV extracted DNA was noted (Josephs et al. 1986). Salahuddin et al. (1986) reported that only 4 of 220'normal' sera and none of 12 anti-HIV positive sera from individuals without lymphomas were reactive against HBLV infected cells by indirect immunofluorescence (IF).

Further isolates of novel lymphotropic human herpesviruses were reported by two laboratories in 1987 (Tedder et al., 1987, Downing et al., 1987). One isolate was obtained from a Gambian AIDS patient infected with HIV-2 (Tedder et al., 1987) by co-cultivation of peripheral blood mononuclear cells (PBMC) from the patient with phytohaemagglutinin (PHA) stimulated cord blood cells, after which the virus was propagated in the J Jhan T-lymphocyte cell line (a clone of the Jurkat cell line). This lymphotropic herpesvirus (LHV, also known as AJ) was found to be closely related to HBLV (as determined by Southern blotting analysis) and no antigenic cross reaction (using specific monoclonal and polyclonal sera) with other human herpesviruses was noted. Cells of B- and T-lymphoid, glial and fibroblastoid origin supported growth of AJ with varying degrees of efficiency. A photomicrograph of the AJ isolate of HHV6 in J Jhan cells is given in Figure 1.1. Preliminary screening of sera from blood donors for antibody to AJ by indirect IF was attempted, 66 sera were examined. Eighteen percent gave strong staining and a further 24 % weaker staining in this assay. The human herpesvirus isolates obtained from HIV-1 infected Ugandan patients with AIDS

28 F i g u r e 1 . 1

Enveloped virus particles in cytoplasmic vesicles

Sections were prepared from HHV6 (AJ) infected J Jhan cells (T-lymphocyte cell line). Bar = 100 nm

(Electron micrograph kindly supplied by Dr D. Robertson, Institute of Cancer Research, The Haddow Laboratories, Surrey)

29 (prototype virus U1102, Downing et al., 1987) were indistinguishable from HBLV and AJ by Southern blotting and again no crossreaction with other human herpesvirus DNA or protein was noted. These Ugandan isolates were able to infect cell lines of T- (CEM, H9, Jurkat) and B-cell origin.

It was confirmed soon after this that the isolates designated HBLV (Salahuddin et al., 1986) were also able to infect cell lines of both T- and B-cell origin (Lusso et al., 1987, Ablashi et al., 1987) and it is predominantly lymphocytes expressing T-cell markers in cultures of peripheral blood mononuclear cells that are susceptible to viral infection (Lusso et al., 1988). It was suggested that the HBLV and Ugandan and Gambian LHV isolates of this new human herpesvirus (together with any subsequent isolates) should be designated human herpesvirus 6 (HHV6) based on the classification of herpesviruses first described by Roizmann et al. (1981).

Since 1987 there has been a number of reports of further novel human herpesviruses isolated from PBMC, tissue samples and saliva (see Tables 4. 1 and 4. 2). Many of these viruses have been characterized and although restriction enzyme pleomorphism has been demonstrated (Josephs et al., 1988, Becker et al., 1989, Jarrett et al., 1989, Kikuta et al., 1989, Levy et al., 1990, Schirmeret al., 1991) most have been confirmed as isolates of HHV6 and are closely related to the original HBLV, U1102 and AJ isolates.

Although it had been confirmed that HHV6 was a new herpesvirus, Efstathiou et al. (1988) demonstrated that a 5.5 Kbp portion of the HHV6 (U1102) genome would cross hybridize with human CMV under high stringency conditions. Nucleotide sequence analysis confirmed that this region of the HHV6 genome was more closely related to CMV than any other human herpesvirus. When more extensive sequencing of a 21.8 Kbp portion of the HHV6 genome covering this region was completed, open reading frames homologous to a set of genes conserved in all other sequenced herpesviruses were identified (Lawrence et al., 1990). Human herpesvirus 6 was found to be as closely related to human CMV across this region as HSV is to VZV (both classified as alphaherpesviruses) and EBV is to HVS (both gammaherpesviruses).

30 lc. Aims of this study and plan of experimental work

The primary aims of this study were to determine the prevalence of HHV6 infection, to examine any disease association of HHV6 infection and to investigate the biological relationship between HHV6 and the other human herpesviruses.

The following plan was developed to achieve these objectives:

1) Development of serological assays for the detection of HHV6-specific antibody.

2) Use of these assays for seroprevalence studies in healthy individuals and assessment of any serological crossreaction between HHV6 and the other human herpesviruses.

3) Identification of patient groups with primary or reactivated HHV6 infection associated with disease using the anti-HHV6 serological assays.

4) Development of techniques for the detection of HHV6-specific DNA and protein in body fluids and tissue samples.

5) Identification of tissue sites and or body fluids involved in viral persistence, transmission and disease association using the HHV6-specific DNA and protein detection methods.

6) Comparison of the epidemiological data obtained for HHV6 with that for the other human herpesviruses in order to assess any potential interactions between them.

7) Assessment of the ancestry (and subsequent classification) of HHV6 using all available data on the molecular and biological relationships between the human herpesviruses.

31 Section 2: Materials and Methods

Unless otherwise stated, disposable plastics used throughout the course of this study were obtained from Flow Laboratories Ltd.. Methods are given in the order in which they appear in the results section and media and buffer recipies are given in Appendix A.

2a. Tissue culture and HHV6 propagation

J Jhan cells (a clonal derivative of the mature T-lymphocyte Jurkat cell line) were cultured in RPMI medium with added supplements (complete RPMI, Appendix A). A cryotube (Nunc supplied by Gibco-BRL Ltd.) containing approximately 3 X 107 J Jhan cells in 1 ml of freezing medium (Appendix A) was removed from liquid nitrogen and thawed in a 37 °C water bath. The cells were washed by addition to 24 ml of RPMI (without supplements) and centrifugation at 150 g for 10 minutes. After a second wash, the J Jhan cells were resuspended in 10 ml of complete RPMI, transferred to a 25 cm2 tissue culture flask and incubated at 37 °C in a 5 % C02 humid atmosphere (GA2 C02 incubator, Leec Ltd.).

After 2 days the concentration of viable cells was determined by counting cells suspended in 0.5 % (w/v) trypan blue (BDH Ltd.) made up in PBS (Appendix A), using an improved Neubauer chamber. The concentration of cells was found to be approximately 3 X 106 cells / ml. The culture was diluted 1 in 3 and transferred to a 75 cm2 tissue culture flask. It soon became apparent that the J Jhan cells remained healthy if split 1 in 3 every 3 -4 days and maintained at a concentration of 1 - 3 X 106 cells / ml in a 75 cm2 tissue culture flask. J Jhan cells were frozen in liquid nitrogen every 2 - 3 weeks. Ten millilitres of a culture due to be split (approximately 3 X 107 cells) were spun out of complete RPMI (150g for 10 minutes) and resuspended in 1 ml freezing medium in a cryotube. The vial was left at - 70 °C overnight before storage in the vapour phase of liquid nitrogen. J Jhan cells stored in this way were successfully revived up to 4 years after they were frozen.

Two other T-cell lines were also cultured during the course of this study. The HSB-2 cell line was derived from immature T-lymphocytes and has been described previously (Ablashi et al, 1987). The c8166 T-cell line carries the partially expressed human T- lymphotropic virus type 1 (HTLV-1) genome and has been shown to support in vitro replication of HIV-1 (Clapham et al., 1987). Both cell lines were cultured and frozen in a similar way to that described for J Jhan cells.

32 The AJ isolate of HHV6 (Tedder et al., 1987) was used throughout the course of this study. Frozen vials of AJ infected J Jhan cells were provided by Miss M. Briggs (Virology Division, University College and Middlesex School of Medicine). The infected cells were revived as described for the J Jhan cells and resuspended in 10 ml of complete RPMI. The culture was monitored by visualization of the CPE by light microscopy and detection of viral antigen by indirect IF. There was very little supernatant virus produced in infected J Jhan cell cultures and so the virus was passaged by transfer of infected cells into uninfected cultures at a ratio of approximately 1 :6.

After passage of virus into uninfected cells the J Jhan cells continued replication for up to 1 week. Three days post infection the newly infected culture was split 1 in 2 and after this time the culture was refed every 3 - 4 days by allowing the cells to settle to the bottom of the flask and replacing half of the supernatant tissue culture medium with fresh complete RPMI. The cultures were examined daily and monitored for viral antigens by indirect IF every 2 - 3 days. Only when > 80 % of the J Jhan cells were infected (after approximately 2 - 3 weeks of culture) was the virus passaged into freshly split J Jhan cells or frozen in the vapour phase of liquid nitrogen. The method used for preparation and storage of infected cell cultures was as described for the J Jhan cells. Infected cells were successfully revived up to 4 years after storage in liquid nitrogen vapour.

In order to passage HHV6 into HSB-2 and c8166 cell lines, AJ infected J Jhan cultures were treated with Mitomycin C to prevent J Jhan cells from taking over the culture. Ten millilitres of the infected culture were pelleted (150 g, 10 minutes) and resuspended in 1 ml of PBS. Freeze dried Mitomycin C (Sigma Chemical Co.) was reconstituted to a concentration of 500 pig / pi and diluted 1 in 6 in the cell suspension. After incubation at 37 °C for 20 minutes the treated cells were washed in PBS and resuspended back into 10 ml of complete medium. This treatment was found to prevent replication of the J Jhan cells in the culture.

After Mitomycin C treatment the infected cells were passaged into HSB-2 and c8166 cells as described for passage of virus in J Jhan cells. Many large syncytia were visible in c8166 cells 3-4 days post infection but it was not possible to passage the virus using supernatants from these cultures. HSB-2 cells were also successfully infected with the AJ isolate with CPE similar to that in J Jhan cells but again no supernatant virus was produced. The AJ isolate of HHV6 was maintained in both c8166 and HSB-2 cell lines by passage of infected cells into freshly split uninfected cells. It was also possible to freeze and revive infected and uninfected HSB 2 and c8166 cells using the methods described.

33 2b. Indirect immunofluorescence (IF):

The indirect IF methods used for the detection of anti-HHV6 were based on those developed for the detection of anti-EBV, which have been described previously (Henle and Henle, 1966, Henle et al ., 1979, Hotchin and Crawford, 1991).

2b. 1. preparation of slides and sera

AJ infected J Jhan cells were used as antigen for indirect IF assays. Infected cultures were harvested when approximately 15 % of cells gave HHV6-specific staining with sera from the individual from whom the HHV6 isolate was obtained. The optimum day post infection to harvest cells for indirect IF was variable, this was probably due to differing amounts of virus in the infected cells which were passaged. Once a panel of sera was available that had been screened by IF, cultures were monitored with negative, weak and strong anti-HHV6 positive sera to ensure the reproducibility of the assay.

To prepare a batch of slides for IF, 30 ml of an infected culture were pelleted (150g, 10 minutes) and washed by resuspending the cells in 30 ml of PBS and centrifuging again. The wash step was then repeated. Finally the cell pellet was resuspended in 3 ml of PBS and 100 pi of the washed cells were added to each well of a teflon coated multiwell slide (C.A. Henley Ltd.). The wells were examined by light microscopy and, when sufficient cells had settled down, excess PBS was removed with a plastic pasteur pipette (BDH Ltd.) and the cell spots were allowed to dry at room temperature. Using this method, approximately 5 X 104 cells were dried onto each well of the multispot slide. The cells were fixed in acetone (99.5 %, BDH Ltd.) at -20 °C for 10 minutes and then stored at - 70 °C in an airtight container until required.

Acetone fixed J Jhan cells were prepared by washing 30 ml of uninfected cells in PBS, as before, and resuspending the cell pellet in 3 ml of acetone. Cells in acetone were left at 4 °C overnight before pelleting and washing twice in excess (30 ml) PBS. Finally, the acetone fixed J Jhan cells were resuspended in 3ml of PBS (concentration of approximately 3 X 107 cells / ml) and stored at 4 °C until required.

All sera to be screened for anti-HHV6 by indirect IF were first absorbed against acetone fixed J Jhan cells. Absorption of sera was usually carried out in 'v ' bottom microtitre plates (Sterilin from Griffiths and Neilson Ltd.). Sera to be tested for anti-HHV6 IgG by indirect IF were diluted 1 in 10 in acetone fixed J Jhan cells (10 pi of serum + 90 |il acetone fixed cells), mixed and left at 4 °C overnight. The cells were then spun out at 50 g for 20 minutes and the absorbed sera stored at - 20 °C until required.

34 Some sera from pregnant women were fractionated by sucrose gradient centrifugation prior to screening for anti-HHV6 IgM by indirect IF. This was to separate the different classes of antibody in order to ensure the fluorescence was associated with the IgM fraction and to avoid false negative reactions due to competing IgG. The method used for fractionation of sera was as described for rubella (Best et al ., 1969, Best and O'Shea, 1989). Linear sucrose gradients were prepared by layering 1.4 ml of 37 %, 23 % and 10 % (w/v) solutions of sucrose (BDH Ltd.) made up in PBS (pH 7.2) on a 0.2 ml cushion of 50 % sucrose in a 5 ml nitrocellulose centrifuge tube (Beckman Instruments Inc). The gradient was left for 4 hours at 4 °C before layering 0.2 ml of the test serum (diluted 1 in 2 in PBS) on the top and centrifuging for 16 hours at 157,000 g in a swing- out rotor (SW 50.1, Beckman Instruments Inc). Fractions were collected by piercing the bottom of the tube with a hot needle. Twelve fractions (each approximately 0.4 ml) were collected and those which contained IgM and IgG were identified using radial immunodiffusion kits (Serotec Ltd.) according to the manufacturer's instructions. Diffusion times were 6 hours for IgG and 48 hours for IgM. Control sera were included on each immunodiffusion plate allowing estimation of the concentration of IgM and IgG in each fraction. The appropriate fractions were screened for anti-HHV6 activity by indirect IF (tested undiluted).

The treatment of sera with RF-absorbans (Behring Diagnostics Ltd.) to remove IgG was more convenient than sucrose gradient fractionation. Anti-HHV6 IgG positive sera gave no precipitation band in a radial immunodiffusion assay against IgG nor any reaction with HHV6 infected J Jhan cells by indirect IF after treatment with this compound. RF- absorbans was also recommended by the manufacturer's for removal of rheumatoid factor (RF) from serum samples. A panel of sera containing RF were screened for anti- HHV6 IgM and IgG by indirect IF before and after treatment with RF-absorbans, results are given in 3a. 7. ii. A latex agglutination test (Rheuma-Wellcotest, Wellcome Diagnostics Ltd.) was used, according to the manufacturer's instructions, for detection of rheumatoid factor in samples used to evaluate the anti-HHV6 IgM assay.

Sera to be screened for anti-HHV6 IgM by indirect IF were absorbed against acetone fixed J Jhan cells at a dilution of 1 in 6 (10 i_l1 of serum + 50 pi of acetone fixed cells), before centrifugation the serum / J Jhan cell suspension was further diluted 1 in 2 in RF- absorbans (made up according to manufacturer's instructions) to remove IgG and anti- IgG. The diluted sera were incubated at room temperature for 15 minutes before centrifugation and storage at - 20°C.

During evaluation of the anti-HHV6 indirect IF assays sera were titrated out in doubling dilution and the fluorescent end-point determined. It was later found that screening sera at a fixed dilution and scoring the strength of the specific fluorescence (+, ++ or +++)

35 gave a good estimate of the anti-HHV6 IgM or IgG titre. A comparison of results using 37 °C and room temperature incubations revealed that the latter gave lower background activity in this assay. Initially, infected cells were incubated for 15 minutes with 25 (il 2 % (v/v) horse serum diluted in PBS but as this first step did not seem to reduce non specific background it was omitted from future assays. The final optimized methods for detection of anti-HHV6 IgG and IgM by indirect IF at a single dilution are given.

2b. 2. detection of anti-HHV6 IgG

Multispot slides with wells containing acetone fixed infected J Jhan cells were thawed and washed once in PBS. Absorbed sera (either fresh or frozen and thawed) were diluted 1 in 5 (final dilution 1 in 50) in PBS and 25 pi of each diluted serum was applied to duplicate wells on the multispot slide. Positive and negative control sera were included on each slide. After incubation in a moist chamber at room temperature for 40 minutes the slides were washed 3 times for 5 minutes in PBS. Fluorescein conjugated anti-human IgG (Dako Ltd.) was diluted in PBS (1 in 100 - 1 in 200, pre-optimized) and Evans' blue (Sigma Chemical Co.) was added to a final dilution of 1 in 1,000, 25 pi of this conjugate was then added to each well. After incubation in a moist chamber for 20 minutes at room temperature the slides were washed in PBS, mounted in PBS / glycerol (1 : 1) and examined using a fluorescent microscope (Carl Zeiss Ltd.). Positive sera were scored +, ++ or +++ depending on the intensity of the fluorescent staining, any background reaction was assessed by comparison with the uninfected cells present in each well. Weakly reactive sera were designated equivocal (±) in this assay.

2b. 3. detection of anti-HHV6 IgM

This assay was similar to that for anti-HHV6 IgG except that sera were pre-absorbed against J Jhan cells and RF-absorbans as described. Sera were further diluted 1 in 2 in PBS (final dilution of 1 in 24) and 25 pi was applied to each well of a multispot slide. After incubation in a moist chamber for 3 hours at room temperature the slides were washed as described for the IgG assay and fluorescein conjugated anti-human IgM (Dako Ltd.) was added (diluted 1 in 40 - 1 in 60 in PBS with Evans' blue). After incubation for 1 hour at room temperature the cells were washed again, mounted and examined. Sera were scored -, ±, +, ++ or +++ depending on the intensity of the fluorescent staining.

2c. Anti-complementary immunofluorescence (ACIF) for detection of anti-HHV6 IgG

An ACIF assay was also developed for the detection of anti-HHV6 based on the method described for detection of anti-EBV nuclear antigen (Reedman and Klein, 1973). The

36 multispot slides prepared for use in the indirect IF assay were also used for ACIF. Sera screened for specific antibody in this assay were heat inactivated by incubation at 56 °C for 30 minutes. Some sera were also preabsorbed against acetone fixed J Jhan cells although this later proved to be unnecessary. The serum used as source of complement in this assay was from an individual who was negative for HHV6 antibody by indirect IF and gave little competition in the radioimmunoassay. All incubations were for 1 hour at room temperature.

Heat inactivated sera diluted in PBS (25 pi of 1 in 20 or 1 in 50) were applied to acetone fixed infected J Jhan cells. After incubation the slides were washed in PBS and the serum used as source of complement was applied (diluted 1 in 10 in PBS). After a further incubation the slides were washed again and 25 pi of fluorescein conjugated anti-C3b (Dako Ltd.) diluted in PBS (1 in 30, pre-optimized) was added. The final incubation was followed by a wash step and the slides were mounted and examined as for indirect IF.

2d. Competitive radioimmunoassay (RIA) for detection of anti-HHV6:

The method used for the development of an anti-HHV6 competitive RIA was adapted from that described for anti-VZV (Tedder et al., 1981). Optimum dilutions of solid- phase antibody, antigen and radiolabel were determined using a 'chessboard' series of dilutions of each constituent in an indirect assay (without competing serum). New batches of globulin, antigen or radiolabelled IgG were also assayed by this method to ensure reproducibility of the assay. Solid-phase antibody (at 4 °C) and antigen preparations (at -20 °C) were stable for at least 6 months. Radiolabelled conjugate was prepared monthly and assayed against antigens prepared from infected and uninfected J Jhan cells to ensure a binding ratio of at least 10 : 1.

2d. 1. globulin preparation

A globulin fraction was prepared by ammonium sulphate precipitation of serum from an individual with a high titre of anti-HHV6 (titre by indirect IF > 1 in 8,000). One millilitre of serum was diluted 1 in 3 in Tris-saline buffer (TSB, diluted from 10X stock, Appendix A), cooled on ice and 6 ml of saturated ammonium sulphate (BDH Ltd., made up in distilled water, pH 7.6) was added with shaking. The diluted serum was left at 4°C overnight and the precipitate spun out at 1,500 g for 10 minutes at 4 °C. The pellet was redissolved in 3 ml of TSB and the ammonium sulphate precipitation was repeated. The final pellet was resuspended in 1 ml of TSB, placed in 3 cm of 8/32" visking tubing (Medicell Ltd.) and dialyzed against 100 ml of TSB overnight at 4 °C. After removal from dialysis tubing the globulin fraction was stored at 4 °C and was stable at this temperature for at least 3 years. Polystyrene Remova-wells (Dynatech

37 Ltd.) were coated with 100 |il of globulin diluted 1 in 1,000 (pre-optimized) in Tris- azide buffer (TAB, diluted from 10X stock, Appendix A) and left for 24 hours at 4 °C. The plates were washed in TAB, each well was filled to the rim with TAB containing bovine serum albumin (Tris-BSA quench, Appendix A) and left at room temperature for 1 hour in order to quench uncoated polystyrene. Excess buffer was removed and the plates were either stored in a moist chamber at 4 °C or used immediately for detection of anti-HHV6.

2d. 2. antigen preparation

When approximately 70 % of cells in an AJ infected J Jhan cell culture were positive by fluorescence 2 X 107 cells were pelleted by centrifugation (150 g for 10 minutes) and washed by resuspending in 30 ml of sterile distilled water and centrifuging again. The wash step was repeated a second time. Finally, the cell pellet was resuspended in 5 ml of distilled water and freeze-thawed rapidly 3 times. The cell suspension was sonicated (20 Kc / s) for 2 minutes and the cellular debris removed by centrifugation (1,000 g for 10 minutes). The supernatant was stored at -20 °C until required.

In a similar way to that described for HHV6 infected J Jhan cells, control antigen was also prepared from uninfected cells in order to assess any non-specific binding of different batches of radiolabelled IgG. This control antigen was also stored at -20 °C.

2d. 3. preparation of 125I-labeIled IgG

An IgG fraction of the high titre anti-HHV6 serum used for coating antibody was prepared by anion-exchange chromatography using DE52 (Whatman Ltd.) as described previously (Tedder and Wilson-Croome, 1980, Sutherland and Briggs, 1983).

One millilitre of serum was dialyzed against 100 ml of 20 mM phosphate buffer (pH 8, diluted from 200 mM stock, Appendix A) overnight at 4 °C. A 10 ml column of DE52 was prepared, allowed to settle for 30 minutes and equilibrated with 20 ml of 20 mM phosphate buffer. The dialyzed serum was allowed to absorb onto the column for 30 minutes before elution with a further 20 ml of phosphate buffer. The effluent from the column was passed through a UV monitor at 280 nm (Pharmacia Ltd.). The conditions used allowed the binding of all serum proteins except IgG to the column and the first peak (IgG fraction) was collected.

The concentration of IgG in the column effluent was estimated using the equation 1.4 OD260nm = 1 m& / ml (1cm light path) and by radial immunodiffusion as described previously. Approximately 3.5 mg of IgG in 2.5 ml of phosphate buffer was obtained

38 from 1 ml of serum in this way. This was stored at - 70 °C until required and was used for up to 4 years after preparation.

Iodination of IgG was by the iodogen method first described by Salacinski et al. (1979). For preparation of iodogen coated tubes, 0.1 mg / ml solution of iodogen (1, 3, 4, 6- tetrachloro-3a, 6a diphenylglycoluril, Pierce and Wariner Ltd.) in chloroform (99.5%, BDH Ltd.) was prepared and 50 pi were added to clean glass test-tubes. The tubes were allowed to dry and were stored in a dessicator at 4 °C until required. Iodogen tubes stored in this way were used for up to 4 years after preparation without reduction in labelling efficiency.

For the iodination reaction, an iodogen tube was placed on ice and 15 pg of the protein to be labelled, diluted to a volume of 80 pi in PBS, was added. The reaction was initiated by addition of 0.5 mCi of sodium 125iodide (125I, Amersham International), the tube was capped and replaced on ice for 10 minutes with occasional gentle mixing.

Separation of 125I-labelled protein from the free iodine was achieved by gel filtration on an 8 ml Sephadex G25 (Pharmacia Ltd.) column. This was prepared and equilibrated with 16 ml Tris-BSA quench prior to initiation of the iodogen reaction. A solution of potassium iodide (BDH Ltd.) in PBS (10 mg / ml) was made up and 0.5 ml was added to the column. The iodination reaction mix was then added, followed by another 0.5 ml of potassium iodide. After absorption of the reaction mix and potassium iodide the column was eluted with Tris-BSA quench. The effluent was passed through a gamma counter (J & P Engineering Ltd.) and the first peak (labelled protein) was collected into 2 ml of 20 % (w/v) BSA (Wilfrid Smith Ltd.) made up in 0.15 M sodium azide (BDH Ltd.). Addition of buffer to the column was continued until the second peak (free label) was eluted. The incorporation of label into the antibody fraction was estimated, usually > 60 % of label was associated with the IgG fraction by this method. Radiolabelled IgG was stored at 4 °C and used for up to 1 month.

2d. 4. detection of anti-HHV6

Initially sera were titrated (2 fold dilutions) in the RIA. It was later shown that the end point was clearly related to the inhibiton of label bound using a single volume of serum. Sera were subsequently screened for anti-HHV6 by competitive RIA using a fixed volume of 5 pi of serum.

Wells coated with antibody and subsequently quenched were aspirated and 100 pi of optimally diluted antigen (1 in 10 to 1 in 30) in TAB / 0.5 % BSA (antigen buffer, Appendix A) were added to each well. Incubation was for 24 - 48 hours in a moist chamber at room temperature, the plates were then washed in TAB. Antigen coated

39 wells were aspirated to dryness and 20 |il of TAB / 2 % BSA (radiolabel buffer, Appendix A) were added to each. Five microlitres of test serum followed by 75 pi of radiolabel buffer containing 100, 000 cpm of 125I-labelled IgG were added to the well, each serum was tested in duplicate wells. The plates were sealed and left in a moist chamber at room temperature overnight, washed in TAB to remove unbound label and the wells counted in a 16-channel gamma counter (NE 1600, NE Technology Ltd.). The maximum labelled antibody bound to the solid phase was calculated for 4 wells containing 5 pi foetal calf serum (FCS, Flow Laboratories Ltd.) instead of the test serum. The average % inhibition of the maximum label bound for each test serum was calculated. Sera from adults giving > 40 % inhibition of label bound were considered to contain anti-HHV6, those giving 30 - 40 % inhibition were designated weakly reactive (±) and those giving < 30 % inhibition were considered not to contain detectable anti- HHV6.

2e. Complement fixation test for detection of anti-HSV, anti-VZV and anti-CMV:

The complement fixation test (CFT) was used for the detection of antibody to HSV, VZV and CMV. Haemolysin (Wellcome Diagnostics Ltd.), complement (Richardson's preserved guinea-pig complement, Wellcome Diagnostics Ltd.) and the viral antigens (Behring Diagnostics Ltd.) were standardized using the appropriate controls according to the methods described (Bradstreet and Taylor, 1962, Schmidt and Emmons, 1989).

2e. 1. Preparation of sensitized sheep red blood cells

Sheep red blood cells in Alsever solution (Difco Laboratories Ltd.) were washed by suspending in 2 volumes of veronal buffer (VB, Appendix A) and centrifuging for 10 minutes at 500 g. The supernatant and top layer of red blood cells were removed and the wash step was repeated twice. The red blood cells were then packed by centrifugation at 700 g for 10 minutes. A 4 % (v/v) solution of packed red blood cells in VB was prepared and stored at 4 °C until required (used within 2 days).

Packed red blood cells were sensitized by addition of an equal volume of haemolysin diluted to the required strength in VB, left at 4°C overnight and then incubated at 37 °C for 30 minutes with occasional gentle mixing.

2e. 2. performance of the test

Test sera were heat inactivated at 56 °C for 30 minutes. Twenty-five microlitre volumes of two-fold dilutions of heat inactivated test sera (1 in 2 to 1 in 256) in VB were prepared in a Lucite microtitre plate. Positive and negative control sera (of known

40 titre) for each antigen were also diluted and included in the assay (supplied by Behring Diagnostics Ltd.). Antigen was diluted in VB and 2 units in a volume of 25 pi were added to each well. Complement was diluted in cold VB and 2 units in a volume of 25 pi were then added to each well. Controls included 1 in 2, 1 in 4 and 1 in 8 dilutions of each test serum in VB with added complement but no antigen to ensure that sera were not anti-complementary. A single well antigen control at the dilution used in the test was also included to check for anti-complementary activity, this well contained antigen and complement but no serum. A complement titration containing 3.00, 1.50 and 0.75 units of complement but no antigen or antibody was prepared to check that the appropriate dilution had been added to the test. One well to be used as the cell control contained only 75 pi VB, all other controls were made up to a volume of 75 pi in VB.

After incubation at 4 °C overnight the plates were warmed at 37 °C for 30 minutes and 25 pi of sensitized sheep red blood cells were added to each well. The plates were shaken and incubated at 37 °C for 15 minutes after which time they were removed, shaken and returned to 37 °C for a further 15 minutes. They were spun (500 g for 10 minutes) and read. Controls were checked to ensure that the red blood cells were not haemolysed in the absence of complement, that the appropriate dilution of complement had been added and that neither sera nor antigens were anti-complementary. Positive and negative control sera were checked to ensure that the appropriate amount of antigen had been added. The titre of test sera in this assay was defined as the reciprocal of the dilution giving 50 % lysis of red blood cells (ie. 50 % complement fixation).

2f. Indirect IF for detection of anti-EBV VCA IgG

An estimate of anti-EBV VCA IgG titre was determined by adaptation of the indirect IF method first described by Henle and Henle (1966).

P3HR1 cells were provided by Dr D. Crawford (Virology Department, Royal Postgraduate Medical School) and were cultured in complete RPMI in a similar way to that described for J Jhan cells. For preparation of slides for IF, 30 ml P3HR1 cells were spun out of medium (150 g for 10 minutes) and washed twice in 30 ml PBS. The cell pellet was finally resuspended in PBS to give a concentration of 1 X 106 cells / ml. As described for anti-HHV6 IF, cells were spotted onto teflon coated multi-well slides and approximately 2.5 X 104 cells were dried onto each well. Slides were fixed in acetone at room temperature for 10 minutes air-dried and stored at -70 °C until required. The method was as described for anti-HHV6 IgG by indirect IF (2b.) except that sera were screened at a fixed dilution of 1 in 200 (diluted in PBS). Incubations with serum and conjugate were for 1 hour at room temperature in a moist chamber. Positive sera were scored +, ++ or +++ depending on the intensity of the fluorescent staining.

41 2g. Solid-phase enzyme-immunoassay for detection of anti-CMV IgM

Anti-CMV IgM was detected using the CMV ELA-IgM kit (Mercia Diagnostics Ltd.) according to the manufacturer's instructions. Plates were supplied pre-coated with anti- IgM, after incubation with the test serum the specificity of bound IgM was determined by addition of peroxidase labelled CMV antigen. The cut-off between positive and negative sera was determined using the borderline, positive and negative controls supplied with the kit.

2h. In situ hybridization for detection of HHV6 DNA:

The in situ hybridization method for the detection of HHV6 DNA was adapted from methods reported previously (Garson et al ., 1987, Anderson et al ., 1988).

2h. 1. preparation of biotin labelled probes by nick translation

A PUC plasmid containing 14 Kbp of HHV6 DNA was provided by Dr R. Honess (National Institute for Medical Research, London). The virus DNA cloned into this plasmid was a Sma I restriction enzyme digest fragment of the U1102 genome (SMD14 fragment, Martin et al ., 1991).

Biotin was incorporated into the plasmid using a nick-translation kit (Amersham International pic. or Gibco-BRL Ltd.) and a biotinylated deoxyribonucleotide triphosphate (biotin-7-dATP, Gibco-BRL Ltd.). The conditions for the nick-translation reaction were as described by the manufacturer of the kit used. A mixture containing 2 nmoles of each unlabelled (dGTP, dTTP and dCTP) and labelled (biotin-7-dATP) deoxyribonucleotide was prepared and added to 2 pg of the HHV6 plasmid (PSMD14) or to 2 pg of the negative control plasmid (PUC8, Gibco-BRL Ltd.) in a 1.5 ml sterile tube (Sarstedt Ltd.). The volume in the tube was made up to 90 pi with sterile distilled water, the contents of the tube were mixed and 5 units of DNA polymerase I and 100 pg DNase were then added in a volume of 10 pi. After mixing again, the tube was incubated at 15 °C for 1.5 hours. The reaction was stopped by addition of 10 pi 0.3 M EDTA (pH 8) and 1.5 pi 10 % SDS.

The plasmid DNA was purified by ethanol precipitation, as described by Sambrook et al. (1989). The reaction volume was made up to 0.4 ml by addition of 80 mmoles of NaCl (BDH Ltd.) in sterile distilled water and 0.8 ml (2 volumes) of ice cold ethanol (BDH Ltd., 99.5 %) was added. After mixing the tube was left at - 20 °C overnight. The DNA was pelleted by centrifugation for 10 minutes at 10,000 g at 4 °C. The supernatant was carefully poured off, the pellet was allowed to dry and then resuspended in 40 pi TE buffer (Appendix A). Precipitation of DNA by this method was assumed to be

42 100 % giving a final concentration of plasmid in TE of 50 ng / pi. Biotinylated plasmids were stable for at least 2 years when stored at - 20 °C.

2h. 2. reproducibility of the nick translation reaction

In order to ensure reproducibility of the in situ hybridization reaction, the incorporation of biotin into each plasmid was determined and compared with that for previous preparations. Dilutions of labelled plasmid equivalent to 10,000, 1,000, 500, 100, 50, 20, 10, 5, and 2.5 pg of DNA in 4 pi of TE were dried onto a nylon membrane (Zetaprobe, Gibco-BRL Ltd.). The tenth spot was 4 pi of TE buffer without plasmid. The membrane was then baked at 80 °C for 2 hours and rehydrated in a Tris based buffer with Triton-X-100 (in situ buffer 1, Appendix A). To reduce any non-specific binding of conjugate and substrate, the membrane was placed in a hybridization bag (Gibco-BRL Ltd.) with 2 ml of a Tris / BSA buffer (in situ buffer 2, Appendix A) and incubated at 60 °C for 1 hour. The filter was removed from the bag and the biotinylated plasmid was reacted with a streptavidin-alkaline phosphatase conjugate (Gibco-BRL Ltd.) diluted 1 in 1,000 (pre-optimized) in in situ buffer 2. Approximately 2 ml of diluted conjugate was used for a 36 cm2 membrane in a square petri-dish and the membrane was incubated in conjugate for 10 minutes, with gentle agitation, at room temperature. The membrane was transferred to a fresh petri-dish and washed with 40 ml of in situ buffer 1 for 10 minutes at room temperature with gentle agitation. The buffer was decanted and the wash step repeated twice with fresh buffer 1 and then once with 40 ml of a high pH Tris buffer (NBT / BCIP buffer, Appendix A). A mix of nitroblue- tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (NBT / BCIP substrate mix, Appendix A) was used for the detection of conjugate bound to the biotinylated DNA. Wash buffer was decanted and 2 ml NBT / BCIP substrate mixture were added. The reaction was allowed to develop in the dark at room temperature for approximately 30 minutes, or until the background reaction (TE spot without DNA) started to become visible. The reaction was terminated by washing the membrane in TE buffer. The membrane was dried in an oven at 80 °C for 2 minutes. Labelled plasmids were only used as probes for in situ hybridization when 5 pg or less of DNA gave a clear coloured reaction (purple) using the method described.

A biotinyated DNA probe specific for CMV DNA was obtained commercially (Enzo Biochemicals Ltd.). The probe was supplied as a mixture of 2 CMV DNA sequences of 22.5 Kbp and 17.2 Kbp which were biotinylated by nick translation. The dilution of this probe which contained an equivalent amount of biotin to the HHV6 and negative control probes was determined as described above. This dilution factor was used to calculate the amount of CMV probe to use in the in situ hybridization reaction.

43 2h. 3. pre-treatment of cells, slides and tissues

Initially, biotinylated probes were reacted with HHV6 infected and uninfected J Jhan cells. Cell spots were prepared by washing cells in PBS as described for the preparation of slides for immunofluorescence (2b. 1.). The final pellet was resuspended in a small volume of PBS and approximately 1 X 106 cells in a volume of 100 pi were spread into a 0.5 cm radius circle on a clean microscope slide (BDH Ltd.). The slides were incubated in Camoy's fixative (Appendix A) for 10 minutes at room temperature and then washed twice in PBS in a 150 ml glass Coplin jar (BDH Ltd.). Each cell spot was covered with 250 pi of 100 pg / ml DNase free RNase (Sigma Chemical Co.) made up in 2X SSC (prepared from 20X stock, Appendix A) and incubated in a moist chamber at 37 °C for 60 minutes. The cell spots were washed again in 2 changes of PBS and dehydrated by transfer into successive glass slide trays containing 50 %, 70 % and 90 % (v/v) ethanol (approximately 2 minutes in each). The cells were then reacted with biotinylated probes as described below.

In order to establish a method for the detection of HHV6 DNA in tissue sections, a pellet of formal-saline fixed infected and uninfected J Jhan cells was resuspended in 2 % (w/v) melted gelatin (BDH Ltd.) and centrifuged at 150 g for 10 minutes. The gelatin was allowed to set and was then removed from the centrifuge tube with the cell pellet intact, embedded in paraffin wax and sectioned as for a piece of tissue. Sectioned infected and uninfected J Jhan cells were used as controls for the in situ hybridization reaction until positive and negative tissue sections were available.

Tissue sections to be screened for HHV6 or CMV DNA were dried onto microscope slides which had been pre-treated with aminoalkylsilane solution. This method was based on that described previously by Rentrop et al. (1986) and was used to ensure that the sections did not lift off the slide when treated with proteolytic enzymes. Microscope slides were cleaned by boiling in 1 % (w/v) SDS (BDH Ltd.) for 10 minutes and rinsing in distilled water and then ethanol. The slides were immersed in fresh 2 % (v/v) aminoalkylsilane solution in dry acetone (Appendix A) for 10 minutes at room temperature, washed in dry acetone for 5 minutes and dried at 42 °C overnight. Treated slides were stored at room temperature until required. Formalin fixed tissues, embedded in paraffin, were processed, sectioned (3|i) and dried onto treated slides by colleagues in Histopathology departments.

Tissue sections were immersed in xylene (BDH Ltd.) at room temperature for 15 minutes to remove paraffin wax and were then rehydrated by successive transfer into glass slide trays containing 100, 90 and 70 % (v/v) ethanol, the slides were then rinsed in distilled water. Initially tissue sections were treated with RNase, as described for cell

44 spots, to ensure that any hybridization signal was due to virus-specific DNA. This treatment did not alter the in situ hybridization results and was later omitted from the final protocol. It was assumed that viral RNA did not survive the fixation, enzyme and heat treatments used during the in situ hybridization reaction.

Dewaxed and rehydrated tissue sections were treated with proteolytic enzymes prior to hybridization. This was found to increase the sensitivity of the reaction by making the tissue more permeable to the DNA probe. Pepsin was found to be suitable for this pre treatment and the protocol used was adapted from methods reported previously (Loning et al., 1986, Bums etal ., 1986, Anderson etal., 1988). A 0.4 % (w/v) solution of pepsin (3,200 units / pg protein, Sigma Chemical Co.) in 0.01 M HC1 was prepared and 100 pi were spread over each tissue section. The sections were then incubated in a moist chamber at 37 °C for between 2 minutes and 45 minutes. The optimum incubation times were 2 minutes for embedded and sectioned tissue culture cells, 20 - 30 minutes for salivary gland tissue and 15-20 minutes for lung, lymph node, spleen and thymus tissue. After pepsin treatment the tissue sections were washed in 3 changes of distilled water (10 minutes / wash) and then dehydrated as described for the cell spots.

2h. 4. hybridization of biotinylated probes to target DNA

Dehydrated cell spots and pepsin treated tissue sections were reacted with biotinylated probe diluted in in situ hybridization mix (Appendix A). Each tissue or cell spot was screened with 50 ng HHV6 (PSMD14), 50 ng negative control (PUC8) or 1 pi of a 1 in 50 dilution of CMV biotinylated probe, each in a volume of 11 pi of hybridization mix. Circular plastic coverslips (16 mm, BDH Ltd.) were placed over sections of tissue or cells and sealed with raw cow-gum applied with a 20 ml syringe. Denaturation of the probe and target DNA was performed in a microwave processor (Bio-Rad H2500) at 80 °C for 15 minutes according to the method of Coates et al. (1987) or in a vacuum oven (Gallenkamp Ltd.) at 90 °C for 15 minutes. The slides were then transferred to a moist chamber in a hot air oven at 42 °C for 18 hours to allow hybridization to occur.

2h. 5. detection of biotinylated probes

After hybridization, the cow-gum was carefully removed and the coverslips floated off the sections or cells by immersion in 2X SSC. The slides were then washed in 2X SSC at room temperature for 30 minutes, 0.1X SSC at 42 °C for 30 minutes and 2X SSC at room temperature for 15 minutes, with gentle shaking of the slide jar during each wash step. Finally slides were immersed in in situ buffer 2 for 10 minutes at room temperature and each section or cell spot was then covered with 100 pi of streptavidin- alkaline phosphatase conjugate diluted 1 in 100 (pre-optimized) in the same buffer. The slides were incubated with conjugate for 20 minutes at room temperature in a moist

45 chamber and then washed in 3 changes of in situ buffer 1 with gentle agitation (5 minutes / wash).

Two different substrates were used for the detection of alkaline phosphatase conjugate bound to the biotinylated probe. For staining with NBT / BCIP, the slides were washed for 10 minutes in NBT / BCIP buffer with gentle agitation and then each section or cell spot was covered with 100 pi of NBT / BCIP substrate. The slides were incubated in a dark, moist chamber at room temperature for 30 - 40 minutes or until the background staining with negative control probe became visible. The reaction was stopped by rinsing the slides in TE buffer and then distilled water. Slides were mounted in PBS / glycerol (1:1) and examined by phase contrast microscopy with a blue filter. The purple / brown staining of cells was permanent and the slides were stored at room temperature after sealing around the coverslips with nail varnish.

A second protocol was developed for detection of alkaline-phosphatase conjugate. After washing in in situ buffer 1, slides were immersed in new fuchsin buffer (Appendix A) for 10 minutes at room temperature and then each section or cell spot was covered with 100 pi of new fuchsin substrate mix (Appendix A) and incubated at room temperature in a dark, moist chamber for 30 minutes or until the background staining with negative control probe became visible. The reaction was terminated by rinsing the slides in distilled water. Sections or cells were counterstained by immersion in haematoxylin (Appendix A) for 10 minutes at room temperature. The slides were then rinsed in distilled water, dehydrated as described previously, cleared by immersion in fresh xylene for 10 minutes, mounted in Canada balsam (Sigma Chemical Co.) and examined by light microscopy. The specific red new fuchsin stain and the purple haematoxylin counterstain were permanent and the slides were stored at room temperature.

2i. Production of monoclonal antibodies against HHV6:

Monoclonal antibodies (MAbs) were prepared against tissue culture derived HHV6 antigens. The methods used were based on those previously reported for other MAbs (Kohler and Milstein, 1975, Tedder et al., 1982).

2i. 1. animals and immunisation protocol

J Jhan cells infected with the AJ isolate of HHV6 were used as source of antigen for preparation of MAbs. Infected cells were washed, freeze-thawed and sonicated as described previously (2d. 2.). Balb/c mice (obtained from Bantin and Kingman Ltd.) were raised in the Animal House in the School of Pathology (University College and Middlesex School of Medicine). A female mouse was injected subcutaneously in the neck with 200 pi of HHV6 antigen which had been emulsified with an equal quantity of

46 Freund's complete adjuvant. Six weeks and 10 weeks after this initial inoculation the mouse was given further 100 pi doses of HHV6 antigen diluted 1 in 2 in PBS. These second and third inoculations were given intraperitoneally. Four weeks after the 3rd inoculation, 100 pi of undiluted HHV6 antigen was injected intravenously into a tail vein. The mouse was killed by C02 3 days after this final inoculation and the spleen was asceptically removed for fusion with mouse myeloma cells.

Tail vein blood was obtained at various stages during the immunisation and assayed for HHV6 antibody by indirect IF. After the third dose of HHV6 antigen, specific antibody was detectable in serum samples from the mouse although the anti-J Jhan cell response was also strong.

2i. 2. culture of myeloma cells

JK-Ag 8653 myeloma cells (Kearney et al ., 1979) were cultured in complete MAb medium (CM, Appendix A) as previously described (Tedder et al ., 1982). The cells were split 1 in 2 twice weekly and maintained at a concentration of 1-2 X 106 cells / ml. The cells were cultured, incubated, frozen and revived as described for J Jhan cells (2a.). In order to ensure that the cells were in the log phase of growth at the time of harvest the cultures were split 1 in 3 24 hours prior to fusion with spleen cells.

On the day of fusion, myeloma cells were spun out (150 g for 10 minutes) and washed twice in the same volume of incomplete MAb medium (IM, Appendix A). The final cell pellet was resuspended in IM at a concentration of 1 X 107 cells / ml.

2i. 3. preparation of feeder cells

Mouse peritoneal exudate cells were used as a feeder cell layer for hybrid cells; these were prepared the day before the fusion was undertaken. Cold CM containing hypoxanthine, aminopterin and thymidine (HAT medium, Appendix A) was used for the preparation of these cells. Five mililitres of this medium was injected into the peritoneal cavity of a freshly killed Balb/c mouse and, after gentle agitation, was removed asceptically now containing exudate cells. The concentration of cells was adjusted to 4 X 106 cells / ml and 100 pi were added to each well of a sterile 96 well tissue culture plate (Northumbria Biologicals Ltd.). The plates were examined on the day of fusion to ensure that that they were free from any visible bacterial or fungal contamination.

2i. 4. preparation of spleen cells

The spleen was carefully removed from the freshly killed, immunised mouse and placed in a sterile plastic petri dish containing 5 ml of cold CM. The spleen was disrupted

47 carefully using blunt forceps and the connective tissue was removed. The spleen cell suspension was transferred to a sterile plastic 'univeraT container (Sterilin Ltd.) and the debris allowed to settle to the bottom. The supernatant was carefully removed and the cells were pelleted at 75 g for 10 minutes. The cells were washed 3 times in 10 ml of IM and the final cell pellet was resuspended in 10 ml of IM. The yield and viability of spleen cells was determined using an improved Neubauer chamber as described previously (2a.), approximately 1.5 X 108 viable spleen cells were obtained.

2i. 5. cell fusion

Myeloma cells (1.5 X 107 cells) were added to the spleen cell suspension and the mixture was centrifuged at 75 g for 10 minutes. The supernatant was removed and the universal container containing the cell pellet placed in a water bath at 37 °C. Polyethylene glycol 1540 (PEG, BDH Ltd.) which had previously been sterilized by autoclaving at a concentration of 50 % (w/v) in IM was also warmed in the 37 °C waterbath. The cell pellet was loosened in the bottom of the universal and 1 ml of the PEG solution was slowly added with gentle mixing. The cell / PEG mix was left at 37 °C for 1 minute and then a further 1 ml of warm IM was added. Further IM was added slowly while keeping the mix at 37 °C. The volume in the universal was approximately doubled every minute until 25 ml of IM had been added. The fused cells were then centrifuged (75 g for 10 minutes) and the pellet resuspended in 50 ml of HAT medium. This cell suspension was distributed evenly over 5 plates of feeder cells (100 pi or approximately 3.3 X 105 cells / well).

2i. 6. culture of hybrid cells

Plates containing fused cells maintained on feeder layers were left in the C 0 2 incubator for 1 week without disturbance. After this time colonies of hybridoma cells were visible by low power microscopy and the cells were refed by replacement of 100 pi of old medium with fresh HAT medium. When the colonies were large enough to be visible by eye (approximately 2 weeks after fusion) the cells were refed with CM containing hypoxanthine and thymidine (HT medium, Appendix A) and the supernatant removed was assayed for anti-HHV6 by indirect IF as described below.

2i. 7. detection of anti-HHV6 secreting hybridomas by indirect IF

Supernatant from wells containing hybridoma colonies was assayed for anti-HHV6 by indirect IF in a similar way to that described for polyclonal serum. Neat supernatant (25 pi) was applied to acetone fixed infected and uninfected J Jhan cells on multiwell slides and incubated for 40 minutes at room temperature. After washing in PBS, 25 pi of fluorescein conjugated rabbit anti-mouse serum (Dako Ltd.) was applied to the slides

48 at a dilution of 1 in 40 (pre-optimized) in PBS. After washing again, the slides were mounted in PBS / glycerol and examined using a fluorescent microscope. Twenty-one parent wells containing colonies secreting anti-HHV6 were chosen for further screening and cloning. The contents of these wells were removed by aspiration and expanded by transfer into 24 well sterile plates (Northumbria Biologicals Ltd.) which had previously been seeded with mouse peritoneal cells in HT medium (approximately 2 X 106 peritoneal cells / well). The final volume of cell suspension in each well was 2 ml; subsequently, lml of supernatant from each well was replaced with fresh medium every 3 - 4 days.

2i. 8. cloning by limiting dilution

Hybridoma cultures in 24 well plates were cloned by limiting dilution once the colonies were large enough to cover the base of each well. For each hybrid culture the volume of cell suspension containing 100 cells was calculated and this volume were added to 10 ml of CM, 100 jj.1 of this cell suspension was added to each of 48 wells of a 96 well plate containing feeder cells in CM. One hundred more cells were added to the remaining 5 ml of CM and 100 pi of this more concentrated cell suspension were added to 24 wells of the plate. Finally, a further 100 cells were added to the remaining cell suspension and this was distributed over the remaining 24 wells (100 pi / well). This variation in the concentration of cells on each plate overcame problems associated with over-dilution due to inaccurate cell counts and low viability of some cultures.

This procedure was undertaken for each of the parent wells selected previously. The cells were refed as necessary by removal of 100 pi of tissue culture supernatant and replacement with fresh CM medium. Wells which contained single hybridoma colonies secreting anti-HHV6 were identified and an average of 3 clones for each original parent well were expanded into 24 well plates in CM as described previously.

2i. 9. culture of hybrid clones

The colonies were refed every 3 days with fresh medium and monitored weekly to ensure the continual secretion of anti-HHV6. Once the colonies were large enough to cover the well they were expanded into 25 cm2 flasks containing feeder cells and finally into 75 cm2 flasks. Once the concentration of hybridoma cells in a culture was high the feeder cells were no longer required. Tissue culture supernatant from each clone was screened by indirect EF aliquoted and stored at - 20 °C until required. Hybridoma cells in freezing medium were frozen in liquid nitrogen vapour as described for J Jhan cells (2a.).

49 The class and sub-class of antibody secreted by hybridomas was determined using a reverse-passive haemagglutination kit (Serotec Ltd.). Red blood cells coated with anti- IgGl5 IgG2a, IgG2b, IgG3, IgA and IgM were supplied and reacted with hybridoma tissue culture supernatants. The kit was used according to the manufacturer’s instructions and the results for different clones from the same parent well were compared.

2j. Immunohistochemical staining for detection of HHV6-specific antigens:

A method was developed for detection of HHV6 antigen by immunohistochemical staining using the avidin-biotin staining techniques first described by Hsu et al. (1981).

2j. 1. pre-treatment of cells, slides and tissues

As for in situ hybridization, infected and uninfected J Jhan cell spots were initially used as controls for the detection of HHV6 antigen by immunohistochemistry. Tissue sections were dried onto aminoalkylsilane treated slides and dewaxed. Cell spots were fixed in Camoy's fixative as described previously (2h. 3.). Sections and cell spots were then treated with 100 pi 0.5% hydrogen peroxide (from 30% stock, BDH Ltd.) in methanol (99 %, BDH Ltd.) in a humidified chamber at room temperature for 10 minutes to block endogenous background activity. Initially, tissue sections and cells were also treated with 100 pi 3 % (v/v) normal swine serum (diluted in TBS) for 15 minutes at room temperature but this proved to be unnecessary for the samples used in this study and was not included in the final protocol. Hydrogen peroxide treated cells and sections were washed in distilled water and then subjected to pepsin treatment as described for in situ hybridization (2h. 3.).

2j. 2. the immunohistochemistry reaction

Pepsin treated sections and cells were washed in distilled water and then 2 changes of TBS (5 minutes / wash) with gentle agitation and covered with 100 pi of the primary MAb. The primary HHV6 antibody was a pool of 5 monoclonal tissue culture supernatants selected because they gave different staining patterns by indirect IF on infected J Jhan cells. Samples were screened for CMV antigen using a commercially available anti-CMV MAb directed against a protein produced late in infection (Dako Ltd., diluted 1 in 50 in TBS). As a negative control, samples were also screened for rubella haemagglutinin using undiluted tissue culture supernatant from an anti-rubella hybridoma culture (Tedder et al., 1982). Incubation with the primary MAb was for 1 hour in a moist chamber at room temperature. The slides were then washed in 2 changes of TBS (5 minutes / wash) and 100 pi of biotinylated rabbit anti-mouse antibody (Dako

50 Ltd.) diluted 1 in 250 in TBS was applied to each cell spot or section. During this incubation (30 minutes at room temperature in a moist chamber) avidin-biotin- peroxidase complex (ABC, Dako Ltd.) was prepared according to the manufacturer’s instructions, this was left to stand for 30 minutes prior to use. The sections or cell spots were washed in 2 changes of TBS, as for the previous step, and 100 pi of ABC complex were added to each. After a 30 minute incubation at room temperature in a humidified chamber the slides were washed in TBS again and covered with 100 pi of diaminobenzidine tetrahydrochloride (DAB) substrate mix (Appendix A). After 30 minutes incubation in a dark, humid chamber or when the background staining (with control monoclonal) started to become apparent the reaction was stopped by rinsing the slides in TBS and then distilled water. The slides were transferred to fresh distilled water and washed with gentle agitation for 5 minutes. Cell spots and sections were then counterstained with haematoxylin, dehydrated, cleared, mounted and examined as described previously (2h. 5.). The DAB stain (brown) and haematoxylin counterstain (purple) were permanent and the slides were stored at room temperature.

2k. The nested polymerase chain reaction (PCR) for detection of HHV6 DNA:

The nested polymerase chain reaction (PCR) was first described by Mullis and Faloona (1987) and is an adaptation of the original PCR method developed for amplification of specific DNA sequences (Saiki et al ., 1985). Details of the experiments to optimize the reaction conditions for HHV6 nested PCR are given in 3c. 1. i and 3c. 1. ii.. The final protocol for HHV6 nested PCR is given in this section.

During the course of this study, nested primers were also prepared for the detection of HSV-, VZV-, CMV- and EBV-specific DNA sequences. The nested PCR method used for detection of these herpesviruses was similar to that described for HHV6.

2k. 1. design and preparation of oligonucleotide primers

A single pair of primers for detection of HHV6 DNA were designed by Dr R. Honess (H6 and H7, Gopal et al., 1990) based on sequence information supplied by Lawrence and colleagues (Lawrence et al., 1990). The 13R region of the HHV6 genome was selected for design of these primers as this area was conserved amongst different isolates of HHV6 (Dr R. Honess, personal communication). A nested pair of oligonucleotide primers was designed which was internal to the first set (nH6 and nH7) for the second round PCR reaction. These internal primers were designed so that both ended in T since the work of Kwok et al. (1990) had shown this to be beneficial in the detection of virus variants with a mismatch at the 3' end of the primer sequence. Primers used for detection of B-globin sequences (PC03 and KM38) have been described

51 previously (Saiki et al ., 1988) and were used in a single round PCR reaction. Details of the design of primers for detection of HSV, VZV, CMV and EB V DNA are beyond the remit of this thesis but the specificity of all oligonucleotide primers used in the nested PCR reaction was confirmed using the Microgenie Data Bank (Beckman Instruments Inc.). Primer sequences and relevant references for detection of HHV6, CMV, HSV, VZV and EBV DNA by nested PCR are given in Table 2. 1.

Oligonucleotide primers were prepared using an Applied Biosystems 381A DNA synthesiser and B-cyanoethyl deoxyribonucleotide phosphoramidites (0.25 pmoles synthesis) according to the manufacturer's instructions. Primers were extracted from a CpG column using fresh 30 % ammonia solution (BDH Ltd.), the process was carried out at 4 °C to minimise any evaporation. Three millilitres of ammonia were drawn into a syringe which was then connected to 1 end of the column containing the newly synthesized primer, an empty 5 ml syringe was attached to the other end of the column. Half a millilitre of the ammonia was passed into the column and left for 15 minutes. This was repeated with successive 0.5 ml portions of the ammonia until all 3 ml had passed into the column. The ammonia was then washed through the column 3 times by passage from one syringe to the other and was finally collected into a glass vial with a teflon coated cap (Wheaton, supplied by Jencons Ltd.). The protective groups were removed from the primer by incubation at 55 °C overnight and the primer in ammonia was stored at - 20 °C.

When required, primers were purified from the ammonia solution by ethanol precipitation in the presence of excess salt (Sambrook et al ., 1989). Usually, 1 ml of ice cold ethanol together with either 80 mmoles of NaCl (from 5 M stock) or 150 pmoles of sodium acetate (BDH Ltd., from 3 M stock at pH 5) were added to 0.5 ml of primer in ammonia. After mixing, this was left at - 20 °C overnight and the DNA pellet spun out and dried as described previously (2h. 1.). The purified primer was resuspended in 100 |il of sterile distilled water and 1 pi of this was diluted 1 in 1,000 in TE buffer for estimation of the primer concentration. This was calculated using the equation 1 OD260 mn = 33 pg / ml single-stranded DNA (1cm light path).

The amount of primer obtained using the method described was approximately 60 % of the theoretical yield. Initially, primers were analysed by FPLC (Pharmacia Ltd.) and using the method described, most of the DNA within the primer preparation was of the required size. There was little variation in purity of different primer preparations and as these primers gave reproducible results when used in nested PCR reactions they were not further purified after ethanol precipitation. Each primer was adjusted to 1 pg / pi with sterile distilled water and stored in aliquots at - 20 °C. Primers (either in ammonia or water) were stable for at least 2 years at - 20 °C.

52 Table 2.1 Primer sequences, anneal temperatures and expected product sizes for herpesvirus DNAs by nested PCR < 73 •2.2 c • P > s3 > 3 7 3 T c 5 1) CO O O o G s e PU, vn m j x £ * * H—» CD W C/3 l-l a c U G O CD - I Si^ 3 l-H S i h * ) X B O t-H a> G & f i t ' c \ O f T « ^ VO C" r~ r~ r- © - t- r~ t-- t-~ - r co o o cs r" o Os fO £ < b° g 3 & 3 8 y iS8 g 5i 3i8 c u < y g § 38 a I £ I M o n i oo

o o o VO (N cm

o o 00 so so s s o O Os o s --1

c3 c bo o c <3 s O Os VO o s O s o C3 «-! csj i so - r —i o o < u < u s o 00 so

& o

s o OQ H U o y 2 U G < U H u U U < O H < < < o < l s b0

o m CO CM - r 1 O <

53 n i m n i n i O m 0s) *—i n i CO i i“ O oo 2 o C O Os Os CO so i 1 - »l t-H »—l |-H —1 < O < E—U O G U O O U o E- U G U G G < < 5= g 8 < O U H 8 G < in r Os rH n i - o VO 00 s O OO VO Q T3

3 co o C 3 c o O O O U H < < s a y u h o o o oo CM in CM J W 03> s Os cn Os n i ^ t i cn t - s 3 - 7 s o VO t-H n 0 0 vo VO O § o o o P O u y £ 8 < o o < < y g o o g u < o o < <1 o in g P r^- d s >n >n s d r^- II SI w 0 OS oo •c m T3

c O co 1 o C cm 8 r~ m Ov CM t"- 3 8 - T h so ^ so s d •—i t-H oo 3" cn - os in oo n i s o t-H U U O G O O H < y o u G uo U 8 < H U G G U < < 8 < g < < vo o cm CM s O s O o < 13 M o 3 3 13 M —i oo CM cn U < H % < U OS oo a • O p M 3" M O 3 o

. > c u U < H b u G G UH GU P U Er< G < * I G < G < in o CO oo CM OS OS o < 3 k* 8 g < < 1

* * ‘ 3 J T3 cxoi XI 1) l+H P-H ^ ) X p •S l-| £ 3 3 * 13 O « O (L> S O co *-• O 3 - x3 x l-H D y C O P P

h 2k. 2. preparation of DNA from tissue, mononuclear cell and oropharyngeal samples

Methods for the extraction of DNA from tissues, peripheral blood, tissue culture cells and oropharyngeal samples were based on those described previously (Sambrook et al.t 1989, Higuchi etal ., 1989, Ausubel etal ., 1989).

Tissue samples (salivary gland or large bowel) were minced and frozen in liquid nitrogen as soon as possible after excision. Approximately 0.2 g of tissue was ground with a mortar and pestle and suspended in 0.5 ml of PCR extraction buffer (EB, Appendix A, Ausubel et al ., 1989) in a 1.5 ml tube.

Mononuclear cells (PBMC) were separated from peripheral blood by density-gradient centrifugation using Ficoll-Paque (Pharmacia Ltd.). Peripheral blood samples were collected into 10 ml vacutainers containing EDTA (Becton Dickinson Ltd.) and spun for 10 minutes at 400 g. The plasma was removed and the cells resuspended in sterile PBS to a final volume of 10 ml. The cell suspension was then layered onto 7.5 ml Ficoll-Paque in a universal container and spun for 30 minutes at 400 g. The mononuclear cells were removed from the PBS / Ficoll-Paque interface in a volume of 2 ml and washed twice by addition of 20 ml sterile PBS and centrifugation at 400 g for 10 minutes. The final cell pellet (approximately 2 X 107 cells) was resuspended in 0.5 ml of EB in a 1.5 ml tube.

Tissue culture cells were spun out of medium and washed twice in sterile PBS as described for peripheral blood samples. Cells from a 10 ml culture (approximately 3 X 107 cells) were resuspended in 0.5 ml EB in a 1.5 ml tube.

Samples of throat washings (TW) were obtained by rinsing the mouth out with sterile distilled water and then with 15 ml of sterile saline (0.14 M NaCl) gargling and rinsing the mouth for 2 minutes. The saline TW sample was ultracentrifuged (100,000 g for 30 minutes) and the pellet resuspended in 0.5 ml EB in a 1.5 ml tube.

All samples in EB were treated in a similar way. They were incubated, at 50 °C for 18 hours and then extracted with an equal volume of phenol / chloroform / isoamyl alcohol which was prepared by mixing buffer saturated phenol (Appendix A), chloroform (BDF1 Ltd., 99.5 %) and isoamyl alcohol (BDF1 Ltd., 99 %) in a ratio of 24 : 24 : 1. After shaking for 10 minutes at room temperature the extracted samples were centrifuged at 14,000 g for 10 minutes and the aqueous (upper) phase transferred to a fresh 1.5 ml tube. The extraction process was repeated once with phenol / chloroform / isoamyl alcohol and once with chloroform / isoamyl alcohol (24 : 1). After the final extraction the aqueous sample was ethanol precipitated, pelleted and resuspended in

54 sterile distilled water as described for primer purification (2k. 1.). The concentration of DNA in extracted samples was estimated using the equation 1 OD260nm = 50 p.g / ml double-stranded DNA (1cm light path).

2k. 3. first round PCR reaction

The reaction conditions first described by Saiki (1989) were used for the development of nested PCR for the detection of HHV6-specific DNA. Throughout the course of this study false positive reactions were avoided by using prevention methods similar to those described by Kwok and Higuchi (1989).

Round 1 of the PCR was carried out in a 50 pi mixture containing 5 pi 1 OX RB (Appendix A), 100 ng (approximately 13 pmoles) of each outer primer (H6 and H7), 10 nmoles of each deoxynucleotide triphosphate (Pharmacia Ltd., supplied as neutralized 100 mM solutions), 1 unit of recombinant Taq polymerase (ILS Ltd.) and 0.1 - 2.0 pg of sample or control DNA. The DNA was added to each tube last and the reaction mixture was overlaid with 100 pi of light mineral oil (Sigma Chemical Co.). Negative controls containing no DNA were included in each PCR reaction. The volume in these tubes was made up to 50 pi with distilled water (added using the same pipette as for sample and control DNAs). After an initial 5 minute denaturation at 94 °C, 35 cycles of 95 °C for 2 minutes (denaturation), 60 °C for 1 minute (annealing of primers) and 72 °C for 1 minute (extension of primer sequence) were carried out followed by a 7 minute extension at 72 °C. This thermal cycling was carried out using an MJ Research programmable heating block (supplied by GRI Ltd.). First round PCR products were held at 4 °C until required.

First round nested PCR for 6-globin, CMV, HSV, VZV and EBV DNA were as described for HHV6 except that the anneal temperature was varied depending on the melt temperature of the primers used. An anneal temperature of 60 °C was used for detection of 6-globin sequences, anneal temperatures used for detection of herpesvirus DNAs are given in Table 2. 1.

2k. 4. second round PCR reaction

The second round PCR reaction mixture was as described for the first round except that 200 ng (approximately 26 pmoles) of each inner primer (nH6 and nH7) was used and 1 pi of the first round PCR product was added as source of DNA. Twenty-five cycles of PCR were sufficient for detection of second round products by agarose gel electrophoresis. Anneal temperatures for second round PCR reactions are given in Table 2. 1.

55 2k. 5. analysis of PCR products

Products of first and second round PCR were analyzed by agarose gel electrophoresis using standard techniques (Sambrook et a l , 1989). A 2 % (w/v) agarose gel containing ethidium bromide was prepared (Appendix A) and allowed to set. Ten microlitres of the PCR product were mixed with 2 p.1 of gel loading buffer (Appendix A) and added to each well. A molecular weight marker was included in each row of 13 sample wells. This marker was 0X174 digested with Hae III (Gibco-BRL Ltd.); 0.5 pi (approximately 0.2 jig) was added per well. Agarose gels were run in TAE (from 50X stock, Appendix A) containing ethidium bromide (gel-running buffer, Appendix A) at 100 volts for 45 minutes. After this time the gels were removed from buffer, the DNA bands examined at 302 nm and photographed using Polaroid 667 film.

21. Statistical methods

Methods used for the statistical analysis of data were as described previously (Parker, 1979, Sokal and Rohlf, 1981, Bland, 1987), 95% confidence limits were employed.

The geometric mean, standard deviation and standard error of the mean for a sample were calculated using standard formulae. Student's t-test was used to compare the means of 2 small samples. The variances of the samples were assumed to be equal and confidence limits for the comparison of means were calculated.

The Chi-squared (X2) test was used to analyse data in a 2 x 2 contingency table where expected values were > 5. The formula for X2 incorporating the Yate's correction was used to allow for non-continuously variable data. Fisher's exact test was used when any of the expected values within a 2 x 2 table were < 5. Both X2 and Fisher's exact tests were 2-tailed.

The G-test of independence (Sokal and Rohlf, 1981) was used for analysis of data in an 8 x 2 contingency table. The formula for the G statistic incorporating Williams' correction was used to allow for small sample size.

The relationship between two variables, plotted as a scattergram, was assessed by calculation of the Spearman rank correlation coefficient using the Apple Macintosh Stats view programme.

56 Section 3: Results

3a. Serological assays for the detection of anti-HHV6

3a. 1. Introduction

Serological assays for the detection of herpesvirus-specific antibody have been used extensively for the identification of primary infection and reactivation. Seroconversion for herpesvirus IgG antibody is considered diagnostic of primary infection but there may also be a demonstrable rise in specific IgG in cases of reactivation and reinfection. Herpesvirus IgM antibody is produced in primary infections and in some cases of reactivation and reinfection. Specific IgM antibody in cord blood samples is indicative of a congenital herpesvirus infection (Griffiths et al., 1982, Longson, 1990).

The CFT has been used widely for the diagnosis of CMV, VZV and HSV infections by demonstration of a 4-fold rise in specific antibody. This method is quick and easy to perform but may not be sensitive enough for the determination of immune status. Immunofluorescence assays have also been used extensively for diagnosis of herpesvirus infections and have the advantage that they may be adapted for detection of different classes of serum antibody. In particular, indirect IF and ACIF assays using EBV-carrying B-lymphocyte cell lines have been refined to allow distinction between IgG responses directed against different EBV antigen complexes. Indirect IF based on the technology first applied to the diagnosis of recent rubella infection (Cradock- Watson et al., 1972) has been modified for the detection of herpesvirus IgM antibody, with some success. A diagnosis of IM is usually made by demonstration of heterophile antibodies in the patient's serum (the Paul-Bunnell test is commonly used) but may be confirmed by detection of anti-EBV VC A IgM by indirect IF (Hotchin and Crawford, 1991).

There have been reports of serological crossreaction between herpesviruses when using CFT and IF for detection of IgG antibody, particularly between those viruses within the same sub-family. Cradock-Watson et al. (1979) used indirect IF and CFT for detection of herpesvirus antibody and reported an apparent increase in anti-VZV IgG in sera from patient's with primary HSV infection. Shiraki et al. (1982) confirmed these results but could not demonstrate crossneutralization between these alphaherpesviruses. Further studies have been undertaken to identify conserved antigens amongst the human herpesviruses which may be responsible for serological crossreaction (discussed in Edson et al., 1985, Balachandran et al., 1987, Heath and Kangro, 1990, Gelb, 1990). Concern over potential serological crossreaction in primary and reactivating herpesvirus

57 infections has led to a great deal of research interest in the further development of sensitive and specific serological assays for the detection of herpesvirus antibody.

Indirect or competitive solid-phase immunoassays are the current methods of choice for the sensitive detection of anti-VZV, HSV and CMV (discussed in Longson, 1990, Heath and Kangro, 1990, Griffiths, 1990). These assays have the advantage of enhanced sensitivity over the CFT allowing a more accurate determination of immune status. Indirect radioimmunoassay (RIA) or enzyme-linked immunosorbent assay (ELISA) are commonly used for detection of VZV, HSV and CMV antibody. The principle of both indirect RIA and indirect ELISA is the same. Viral antigen is fixed to a solid-phase (usually via an antibody carrier) and is incubated with the test serum. After washing away unbound antibody the virus-specific IgG is detected using a radiolabelled anti-IgG conjugate (RIA) or an enzyme labelled anti-IgG conjugate with a substrate specific for the enzyme (ELISA). One potential disadvantage of RIA or ELISA for detection of IgG antibody is that pure, standardized reagents are needed to ensure the specificity of these very sensitive assays (Kangro, 1990). A competitive format immunoassay in which the test serum competes with radiolabelled or enzyme labelled virus-specific IgG may overcome some of the potential problems of serological crossreaction. A competitive immunoassay (using a radiolabelled or enzyme labelled IgG conjugate) for detection of anti-VZV has been reported (Tedder et al , 1981). This assay was used successfully for the identification of VZV antibody positive and negative individuals and ^as not affected by the presence or absence of antibody to the other human herpesviruses. An ELISA has been developed for detection of anti-EBV using purified tissue culture derived antigens but as IF has proved so successful for detection of EBV antibody this assay has not been widely used (Hotchin and Crawford, 1991).

Indirect solid-phase immunoassays may be adapted for the detection of IgM antibody by using an anti-IgM radiolabelled or enzyme labelled conjugate. In particular, an indirect RIA has been used successfully for the detection of anti-rubella IgM (Kangro et al ., 1978). A potential problem with the use of indirect immunoassays for the detection of IgM antibody is that the presence of rheumatoid factor (RF) may give a false positive reaction. This phenomenon has been reported for some anti-CMV IgG positive sera containing RF that were screened for CMV-specific IgM by indirect RIA (Kangro, 1980). A false negative reaction may also occur if there are significant levels of specific IgG in the test sample. Antibody capture assays developed for the detection of anti hepatitis B core IgM and anti-rubella IgM (Tedder and Wilson-Croome 1980, Mortimer et al., 1981) have since been applied to the detection of anti-VZV, HSV and CMV IgM. In this assay antibody to the Fc portion of IgM is used to 'capture' antibody from the test serum. Viral antigen is then added and after washing this is detected using a radiolabelled or enzyme labelled virus-specific antibody. The VZV IgM capture RIA

58 has been particularly successful in the identification of primary VZV infection and specific IgM antibody has also been detected in samples from most zoster patients (Tedder et al., 1981, Kangro et al., 1988), similar assays for the detection of HSV IgM are prone to non-specific reactions and are not as yet in general use (Longson, 1990). Indirect RIAs and IgM capture assays have been used to confirm congenital and other primary CMV infections (Knez et al., 1976, Griffiths et al ., 1982, Sutherland and Briggs, 1983, Griffiths and Kangro, 1984) and CMV IgM antibody has been detected in a few cases of CMV reactivation in renal transplant patients (Sutherland and Briggs, 1983). An IgM capture ELISA utilizing nuclei from an EBV producer cell line for diagnosis of IM has been described (Wielaard et al., 1988) but has not been widely used. Detection of anti-EBV VCA IgM by indirect IF is most commonly used for confirmation of primary EBV infection (Hotchin and Crawford, 1991).

The IF and RIA techniques that have proved useful in the identification of herpesvirus infections were applied to the detection of HHV6 specific antibody. This sub-section describes the optimization of these anti-HHV6 serological assays and their use in seroprevalence studies.

3a. 2. Detection of anti-HHV6 IgG by IF:

The AJ isolate of HHV6 was successfully cultivated in the J Jhan T lymphocyte cell line as described previously (2a.). Although the AJ isolate was found to grow in a wide variety of lymphocytic cell lines and cord blood cells, infected J Jhan cells gave the most reproducible results in IF assays. After passage of HHV6 infected cells into freshly split J Jhan cells (ratio infected : uninfected, 1 : 6) swollen cells and syncytia were visible after 3-5 days as shown in the photomicrographs Figure 3a. 1 and Figure 3a. 2. Cells were usually harvested for IF when 15-20 % of cells were giving HHV6 specific staining.

3a. 2. i. titres and scoring system for indirect IF

Serum samples from 96 blood donors were screened for anti-HHV6 IgG by IF at a fixed dilution of 1 in 50 and sera eliciting fluorescence scored +, ++ or +++ depending on the intensity of the staining. The background of uninfected cells (80-85 %) was used to assess any non-specific reactivity. A number of sera were designated equivocal (±) at the screening dilution, these sera gave distinct, but dull, staining of the right proportion of infected cells. Five blood donor sera from each group scored negative, ±, +, ++ and +++, respectively were titrated in doubling dilution (1 in 10 to 1 in 1280) and the titre given as the last dilution with significant fluorescence. A comparison of anti-HHV6 IgG score at 1 in 50 and titre for 25 blood donors is given in Figure 3a. 3. Sera designated

59 Figure 3a. 1 HHV6 CPE in the J Jhan T-cell line

Against a background of normal cells not yet showing CPE, swollen cells, early clear multinucleated cells and later very granular syncytia are seen (x 400).

Figure 3a. 2 Section through a pellet of HHV6 infected J Jhan cells

* « .

A multinucleated syncytium is shown, cells are stained with haematoxylin (x 400).

60 Figure 3a. 3

A comparison of anti-HHV6 IgG score and titre by indirect IF for 25 blood donor sera

+++ ♦ ♦ ♦ O m

++ ♦ ♦ ♦

UQi © ♦ ♦ Vi o i«—i60 '© ♦ ♦ ♦ > as x Neg ♦ ♦ ♦

<40 ^40 40 80 160 320 640 1280 HHV6 IgG titre

rank correlation coefficient = 0.99 (p < 0.001)

Table 3a. 1

Range of titres for each group of sera scored for anti-HHV6 IgG by indirect IF

IF score at IF 1 in 50 titre

Neg < 4 0

+ ^ 4 0

+ 40-80

+ + 160

+ + + ^ 320

61 negative or ± for anti-HHV6 IgG by IF at 1 in 50 did not give unequivocal staining when screened at a dilution of 1 in 10 or 1 in 20. The score at 1 in 50 was clearly related to the end point titre for anti-HHV6 IgG by IF in this group (rank correlation coefficient 0.99, p < 0.001) and the scoring system was adopted for antibody prevalence studies. The range of titres for each scoring group are given in Table 3a. 1. Photomicrographs of the fluorescence produced by sera designated negative, ±, +, ++ and +++ at the screening dilution are given in Figure 3 a. 4.

3a. 2. ii. reproducibility of indirect IF

Care was taken to avoid variation in the slides prepared for IF as described previously (2b.). Each batch of slides was tested with a set of 12 reference sera which had been screened previously and scored. While optimizing the indirect IF test for anti-HHV6 IgG, sera were tested in duplicate on at least 2 separate occasions and the fluorescence was read and scored independently by three different laboratory workers. Once the appropriate scoring system had been established and reproducibility of the test confirmed, samples were tested at a single dilution (1 in 50) and read and scored by myself with reference to control sera on each slide. Sera were retested and the score confirmed on a second occasion.

3a. 2. iii. comparison of indirect IF and ACIF for detection of anti-HHV6 IgG

A panel of 40 blood donor sera which had been screened by indirect IF were subsequently tested by the anti-complementary IF (ACIF) method for anti-HHV6 IgG. Heat inactivated sera were screened at 1 in 20 and 1 in 50 by this method using an indirect IF negative serum as source of complement. The staining by ACIF was more discrete than that for the indirect method and there was very little background staining, even when testing unabsorbed sera at 1 in 20. Photomicrographs of the fluorescence produced by an anti-HHV6 IgG positive serum by indirect IF and ACIF are given in Figure 3a. 5. Sera which were equivocal (±) in the indirect test were not resolved by ACIF. The sera that were scored negative by indirect IF were negative by ACIF and sera that were scored +, ++ or +++ by indirect IF were also positive by ACIF. As there appeared to be no difference in the sensitivity of the two IF techniques the indirect method, which was simpler and less time consuming than ACIF, was employed for further serological studies. Sera designated equivocal were not included when quoting the seroprevalence of HHV6 IgG antibody by IF. This probably led to an underestimation of the total number of HHV6 infected individuals since at least some of these sera would most likely be confirmed as genuinely positive by a more sensitive serological assay.

62 Photomicrographs of indirect IF screening of sera for anti-HHV6 IgG

+ * f f ^ l

4 # 1 »

f W

+ + + + +

The fluorescent staining of acetone fixed HHV6 infected J Jhan cells by anti-HHV6 IgG positive sera is both nuclear and cytoplasmic. Sera were diluted 1 in 50 and cells were counterstained with Evans' blue (x 1600).

63 Photomicrographs of the fluorescence produced by an anti-HH\ 6 IgG positive serum by ACIF and indirect IF

HHV6 specific staining by ACIF (A) is more discrete than that by indirect IF (B). The serum dilution was 1 in 50 (x 1600).

64 3a. 3. Detection of anti-HHV6 by competitive RIA:

Antigen prepared from HHV6 infected J Jhan cell lysates was found to be suitable for use in an RIA for detection of anti-HHV6. Initially, 125I-labelled conjugates were prepared from a number of HHV6-specific IgG antibody positive sera which had been identified by IF, but the discrimination between anti-HHV6 and anti-J Jhan reactions was not good. It was only when an individual who had reactivated HHV6 and had very high levels of specific IgG antibody was identified that the specificity of the test was enhanced (Patient 21, see 3b. 2.).

3a. 3. i. optimization of the assay

The optimum dilution of solid-phase IgG antibody and HHV6 antigen was determined using a 'chessboard" series of dilutions in an RIA without competing serum. Maximum specific binding of the radiolabelled conjugate with the HHV6 antigen was achieved using a 1 in 1,000 dilution of solid-phase IgG and a 1 in 10 to 1 in 30 dilution of HHV6 cell lysate. The average binding of the label to HHV6 antigen was approximately 2 % (2,000 cpm) compared with 0.13 % (133 cpm) binding to control antigen. To ensure maximum specificity, the optimized reagents were then used in a competitive format RIA for the detection of anti-HHV6. Sera were tested undiluted in this assay. When setting up the competitive RIA the label bound in the presence of normal rabbit serum, goat serum, horse serum and FCS was compared. There was little difference in the label bound in the presence of any of these animal sera (Table 3a. 2) and FCS was used as a 'negative" in subsequent experiments to give a value for maximum binding of label. Approximately 2 % (2,000 cpm) of the labelled IgG bound to the solid phase when there was no specific competing antibody present compared with 0.09% (90 cpm) binding of label in the presence of high titre anti-HHV6.

3a. 3. ii. titration

Serum samples from 96 blood donors were tested undiluted for anti-HHV6 IgG by competitive RIA. The average % inhibition of the maximum expected label bound by each test serum was calculated. The sera were divided into 5 groups (0-20 %, 21-40 %, 41-60 %, 61-80 % and 81-100 % inhibition, respectively) and five sera from each group titrated out in the competitive RIA (doubling dilutions from 1 in 2 to 1 in 31,622 in FCS). The average titration results for the 5 categories of sera are given in Figure 3a. 6. The best discrimination between the groups was obtained by testing the sera undiluted and the % inhibition of undiluted serum was clearly related to the titre. The most divergent results were for those sera giving 81-100 % inhibition when tested undiluted. Titration results for these 5 sera are given separately (Figure 3a. 7.).

65 Table 3a. 2

Radiolabel bound to HHV6 antigen in the presence of different animal sera

Mean label bound ± SD Serum for 8 duplicates (cpm)

goat 2083 ± 140

horse 1996 ± 79

rabbit 2 1 4 2 ± 106

FCS 2011 ± 99

Five microlitres of each serum was competed with 100,000 cpm of radiolabelled anti-HHV6. Average label bound in the presence of high titre unlabelled anti-HHV6 was 88 (± 9) cpm for this assay.

66 Figure 3a, 6 3a, Figure -o— sera scoring 0- 20 % inhibition when tested undiluted tested when undiluted % inhibition 20 tested 0- when % inhibition undiluted tested 1-40 2 when scoring sera % inhibition 0 scoring -6 1 4 sera -o— scoring sera *— *— ■m ± 3 -18% inhibition. For clarity, % inhibitions below below inhibitions % clarity, For inhibition. -18% 3 ± 5 % are not shown. not are % 5 between varied point sera, 5 each for for results deviation titration standard the average represents point Each — sera scoring 81-100 % inhibition when tested undiluted tested when inhibition % 81-100 scoring sera — — sera scoring 6 1 -8 0 % inhibition when tested undiluted tested when % inhibition 0 -8 1 6 scoring sera — % inhibition of label bound Titration of 5 categories of sera in the in sera of categories 5 of Titration 100 n 100 701 90 i 90 40- 60- 30- 50- 80- 0 1 - 0

anti-HHV6 competitive RIA competitive anti-HHV6 1

_ __ 3 □

’—o serum dilution (Iog2) dilution serum 4

67 7

15 Figure 3a. 7 Titration of 5 sera in the anti-HHV6 competitive RIA: assessment of variation between duplicate reactions in one assay o o punoq pquj jo uoiqqiqui % uoiqqiqui jo pquj punoq % uoqiqiquj jo pqu[ punoq uo j uqqqi % uoqiqiqui jo q q p punoq o o o o o O VO VO o o o O

■3 ■3 ■3 "O o o o 3

punoq pquj jo uoqiqiqui % uoqiqiqui jo pquj punoq punoq p q q jo uoqiqiqui % uoqiqiqui jo q q p punoq vO o vO XJ 'H

■3 3 3a. 3. iii. reproducibility

All preparations of solid-phase IgG and antigen were diluted and assayed without a competing test serum to ensure reproducibility of the RIA and to confirm a binding ratio of at least 10:1 (HHV6 antigen : control antigen). Once the reproducibility of the RIA had been confirmed the result for each test serum was calculated as the average % inhibition of the maximum expected label bound from duplicate wells, the result was checked by retesting the serum on a second occasion. The results for duplicate wells were very consistent as shown for 5 HHV6 antibody positive sera titrated in the anti- HHV6 competitive RIA (Figure 3a. 7). A panel of control sera which had been titrated previously were included in every RIA plate to ensure there was no plate to plate variation. Results for these 5 control sera tested for HHV6 antibody in 12 consecutive assays are given in Figure 3a. 8. The strong positive control serum used was from the same individual from whom the radiolabelled conjugate was prepared (serum 1, Figure 3a. 8). The average ratio of fully inhibited cpm : uninhibited cpm (serum 1 : FCS ) for the 12 consecutive assays was 1 : 22 (range 1 : 20 - 1 : 26) and this ratio was used as a guideline to determine the success of future assays.

3a. 4. Anti-HHV6 in serum samples from blood donors:

3a. 4. i. by indirect IF

The results for anti-HHV6 IgG by indirect IF in serum samples from 96 blood donors are given in Figure 3a. 9. Approximately 60 % of blood donor samples were scored +, ++ or +++ in this assay and designated positive for anti-HHV6 IgG. The discrimination between a weak positive (+) and negative in this group was not good with 23 % of samples giving equivocal results by indirect IF.

The age and sex was known for 93 of the blood donors and the results for HHV6 antibody in this group, divided into male and female, are given in Table 3a. 3. The distribution of high level reactivity by IF was significantly different when comparing males and females; 8/50 males compared with 18 / 43 females scored ++ or +++ in this assay (X2 test, p < 0.05). A comparison of the age of donor with the score for HHV6 IgG by IF is given in Figure 3a. 10, there was no correlation between age and score by IF for male and female blood donors.

3a. 4. ii. by competitive RIA

The results for anti-HHV6 by competitive RIA in serum samples from 96 blood donors are given in Figure 3a. 11. The histogram shows some evidence of bimodality

69 Figure 3a. 8 3a. Figure

% inhibition of label bound ah eu r ie «>- ). («-><-« given are serum each of each serum in a single assay. Results for 12 consecutive consecutive 12 for Results assay. single a in serum each of Each point represents an average average an represents point Each assays are shown and the mean mean the and shown are assays lOO-i Results for 5 control sera screened for screened sera control 5 for Results assessment of inter-test variability inter-test of assessment RIA: competitive by anti-HHV6 20 H20 0- .... - 40 70- H 90 0- ill l i -I 50 60- 0i * - „ *§ J- i 30 80- 10-1 I J 0 ------4 3 2 1 I

I 1 ------♦ -L ♦♦ t T I 1 70 serum ------t T % % inhibition and SD for for SD and inhibition I 1 inhibition for duplicates duplicates for inhibition ------X * T ♦ H 1

1 l : T X

Figure 3a. 9 3a. Figure Number of Blood Donors Anti-HHV6 IgG by indirect IF in sera in IF indirect by IgG Anti-HHV6 40- 50" 60- 30- from 96 blood donors blood 96 from F t- ( i 5 ) 50 in 1 ( 6 V H nti-H A IF 71 Table 3a. 3

Anti-HHV6 IgG by indirect IF in sera from male and female blood donors

n = 93 Neg + + + + + + + Strong Positive

12 11 19 7 1 Male n = 50 (24%) (22%) (38%) (14%) (2%)

5 10 10 12 6 Female (12%) n = 43 (23%) (23%) (28%) (14%)

total positive male sera 27/50 = 54% strong positive male sera 8/50 = 16% total positive female sera 28/43 =65% strong positive female sera 18/43 = 42%

The distribution of anti-HHV6 high level reactivity by IF (++ and +++) was significantly different when sera from males and females were compared (8/50 compared with 18/43, X2 test, p < 0.05). Figure 3a. 10

Anti-HHV6 IgG score by IF compared with age for 93 blood donor sera

+++ o o o • o o o

© ++ • • • • • • in • o o o o 0 o o 0 o o 0 o c • • • • tt- + • • • • • • • • • • • • • • o o o 0 ° 0 o • 0 o o 'sO > X • • • X • • • • i + O 0 o o 0 • 0 • • O O • 0 o C c - • • • • • Neg • • • • o • • o o o o • 1 1 1 10-20 21-30 31-40 41-50 51-60 >60

Age (years)

• male blood donor sera (n= 50) ° female blood donor sera (n= 43)

There was no significant correlation between age of blood donor and score for anti-HHV6 IgG by indirect IF; rank correlation coefficients -0.10 (females, p > 0.1) and 0.06 (males, p > 0.1).

73 iue3. 11 3a. Figure Number Anti-HHV6 by competitive RIA in sera in RIA competitive by Anti-HHV6 0 i 20 Table3a. 4 10 - 5 - 5 - -0 12 2-0 14 4-0 16 6-0 18 8-0 91-100 81-90 71-80 61-70 51-60 41-30 31-40 21-30 11-20 0-10 Scoringsystem foranti-HHV6 from 96 blood donors blood 96 from bycompetitive RIA % inhibition of label bound label of inhibition % Score + + + + + Neg ++ + + r > ***** i > » r » 74 flbl bound label of % inhibition inhibition % 81-100 41-60 61-80 31-40 0-30 indicating that the blood donor sera may be divided into those with high level anti- HHV6 and those with only a low level of specific antibody. The discrimination between positive and negative by competitive RIA was not clear and it was difficult to establish true negatives in this group (22 % of samples gave < 40 % inhibition). The competitive RIA was optimized using only 5 pi of test serum so that the small amounts of serum obtained from children could also be screened for anti-HHV6 by this assay. It is possible that the discrimination between positive and negative in the blood donor group would have been clearer had more test serum been used per assay. Alternatively all adult sera may contain anti-HHV6. All of the sera from blood donors tested in this assay gave some inhibition of label binding when compared with FCS. Sera taken from children before seroconversion for HHV6 (see 3b. 1.) were considered to be genuinely negative for anti-HHV6 although they may have contained low levels of maternal antibody. These sera gave inhibitions varying from 16-27 % in the competitive RIA. Based on these results a scoring system for anti-HHV6 by RIA was developed (Table 3a. 4) and used for the estimation of antibody prevalence in male and female blood donors. Serum samples from adults giving > 40 % inhibition of label binding were considered anti-HHV6 positive, those giving 31-40 % inhibition were designated equivocal (±) and those giving < 30 % inhibition were classified as negative in this assay. Sera designated equivocal were not included when quoting the seroprevalence of anti-HHV6 by competitive RIA. This probably led to an underestimation of the total number of HHV6 infected individuals. Approximately 80 % of blood donors were designated +, ++ or +++ in this assay and considered to be anti-HHV6 positive, a further 10 % were classified as equivocal. Using this scoring system the 5 control sera tested in each RIA were designated +++, ++, +, ± and negative, respectively. The ratio of fully inhibited cpm : borderline cpm (serum 1 : serum 4, Figure 3a. 8) was consistently in the range 1 : 12 - 1 : 17.

As before, the blood donors were divided into males and females and the competitive RIA results for these groups are given in Table 3a. 5. There was no significant difference in the number of strong positive (> 60% inhibition, ++ and +++) sera when comparing male and female blood donors. There was also no correlation between age and score by competitive RIA for male or female blood donors as shown in Figure 3a. 12.

3a. 4. iii. correlation between indirect IF and competitive RIA

A comparison of HHV6 antibody levels by indirect IF and competitive RIA was made using the results for the blood donor group (Figure 3a. 13). There was good correlation between score by IF and % inhibition by competitive RIA (rank correlation coefficient 0.78, p < 0.001). All but one of the sera scored +, ++ or +++ by IF gave >40 %

75 Table 3a. 5

Anti-HHV6 by competitive RIA in sera from male and female blood donors

% inhibition and score n = 93 0-30 31-40 41-60 61-80 1 81-100 Neg + + ++ ' +++ Strong Positive

5 7 14 14 10 Male n = 50 (10% ) (14% ) (28% ) (28% ) (20% )

6 2 13 11 11 Female n = 43 (14% ) (5% ) (30% ) (26% ) (26% )

total positive male sera 38/50 = 76% strong positive male sera 24/50 = 48% total positive female sera 35/43 = 81% strong positive female sera 22/43 = 51%

There was no significant difference in the distribution of anti-HHV6 high level reactivity (++ or +++, > 60% inhibition) by RIA when sera from males and females were compared (24/50 compared with 22/43, X2 test, p > 0.1).

76 ° female blood donor sera (n= 43) (n= sera donor blood female ° • male blood donor sera (n= 50) 50) (n= sera donor blood male • iue3. 12 3a. Figure There was no significant correlation between age of blood donor donor blood of age between correlation significant no was There coefficients -0.18 (females, p > 0.1) and -0.14 (males, p > 0.1). > p (males, -0.14 and correlation 0.1) > p rank RIA; (females, -0.18 competitive by coefficients anti-HHV6 for score and

Anti-HHV6 RIA Score compared RIA competitive by score Anti-HHV6 + + Neg + + + + + + + with age for 93 blood donor sera donor blood 93 for age with 02 21-30 10-20 • • OO O • O • O • • • 0 • • • • •• • • • • • O OO o • • o 0 • • • • • • O 0 o • o • • • o • o O0 • • • 0 • •

o O O •

14 41-50 31-40

Age (years) Age o • 0 • • • • o •

o OO • • • O O • • 0 O 0 0 o o o o o O • • O • O 51-60 o • o o O • 0 • • 1 o >60

iue3. 13 3a. Figure Anti-HHV6 IgG by IF (1 in 50) + + + + + + + + Correlation between indirect IF and IF indirect between Correlation 0-10 rank correlation coefficient 0.78 (p < 0.001) < (p 0.78 coefficient correlation rank competitive RIA for anti-HHV6 for RIA competitive 11-20 in sera from 96 blood donors blood 96 from sera in ♦ ♦ ♦♦ ♦ ♦♦ ♦♦♦ ♦♦♦ ♦♦♦ ♦♦♦ ♦♦♦ ♦♦

21-30 Anti-HHV6 by competitive RIA competitive by Anti-HHV6

31-40 % inhibition) (%

41-50 ♦ ♦♦ ♦♦♦ ♦♦♦ ♦♦♦ ♦ ♦ ♦ ♦ ♦ ♦♦ ♦ ♦ 78

51-60

61-70

♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ 71-80 ♦♦♦ ♦ ♦♦♦ ♦

♦♦♦ ♦♦♦ 81-90

91-100

inhibition of label binding in the competitive RIA. More than half of the samples designated negative or equivocal by IF were positive by RIA suggesting that the latter method is more sensitive for the detection of anti-HHV6.

The greater sensitivity of the RIA for detection of anti-HHV6 was confirmed when a group of 25 cord blood samples were tested for HHV6 antibody. Only 24 % were HHV6 antibody positive by indirect IF compared with 80 % positive by RIA.

3a. 5. Anti-HHV6 IgG in serum samples from children

Serum samples from 252 children aged 1-17 years were tested for anti-HHV6 IgG by indirect IF (Table 3a. 6). The prevalence of HHV6 antibody seemed constant throughout this age range but in 112 sera from infants the prevalence of anti-HHV6 IgG declined over the first 6 months of life (Table 3a. 7). From age 6 months the prevalence rose and by 1 year of age the HHV6 antibody levels by indirect IF were similar to those found in samples from blood donors (63 % compared with 58 %, respectively). There was not sufficient serum from these children to test for anti-HHV6 by competitive RIA.

3a. 6. Comparison of HHV6 antibody and antibody to the other human herpesviruses in sera from adults and children

Sera from the 96 blood donors were also tested for antibody to the other human herpesviruses. Antibodies to HSV, CMV and VZV were measured by a standard CFT. Anti-EBV VCA IgG was measured by indirect IF at a fixed dilution using a scoring system similar to that for the anti-HHV6 IF test. Sera from 37 blood donors were titrated and tested for anti-VC A IgG by laboratory workers at the Royal Postgraduate Medical School. A comparison of score at 1 in 200 with titre is given in Figure 3a. 14. The score at a fixed dilution was clearly related to the end point for EBV VCA IgG (rank correlation coefficient 0.47, p < 0.005). The level of antibody to HHV6 and to each of the other herpesviruses was compared in this group of blood donors and the results are given in Figures 3a. 15 and 3a. 16. There was no significant correlation between the presence or level of HHV6 antibody (as measured by indirect IF or competitive RIA) and presence or level of antibody to any of the other herpesviruses in sera from these blood donors.

There was sufficient serum from 28 children (mean age 22 months, range 1-97 months) to test for antibody to the other human herpesviruses. A comparison of antibody to HHV6 (by indirect IF) and antibody to each of the other human herpesviruses is given in Figure 3a. 17. There was no significant correlation between antibody to HHV6 and antibody to any of the other human herpesviruses in this group.

79 Table 3a. VO 2 O £ f HH —« c ^ £ < 3 -o ■3 ^ fa a CD© , CD© ^ c / eg c/3 u o> U eg © , C/3 fs • eg a/ mm

80 (N cn VO lO ^+ oc r~- + + + +

Age (years) 9+10 11+12 13+14 15-17 m \ - n S \ S 2 — m i—\ \ cn m rn CO \ o l ~ No. M c \ ~

14 ^ \ 25 Positive^ V \

Anti-HHV6 IgG by indirect IF in sera from 112 children under 1 year of age

Age (months) < 2 2+3 4+5 6+7 8+9 10+11

No. Positive^ 9 7 1 / No. / X / S' s ' S' Tested 22 23 14 22 15 16

% Positive 41 30 7 45 47 63

81 iue3. 14 3a. Figure Anti-EBV score (1 in 200) + + + + + + + + + + ♦ ♦ ♦♦ 1 1 2 4 8 10 2 60 1280 640 320 160 80 40 20 10 <10 Anti-EBV VCA IgG by indirect IF indirect by IgG VCA Anti-EBV Comparison of score and titre for titre and score of Comparison rank correlation coefficient 0.47 (p < 0.005) < (p 0.47 coefficient rankcorrelation in 37 blood donor sera donor blood 37 in Anti-EBV titre Anti-EBV 82 ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦♦♦ ♦ ♦ ♦

Figure 3a. 153a.Figure

Anti-VZVCF Titre Anti-HSV CF Titre the to antibody and IF by IgG anti-HHV6 of Comparison 128 < 128 32 64 <2 16 32 64 16 4 8 2 2 4 8 2 other human herpesviruses in 96 blood donor sera donor blood 96 in herpesviruses human other 0.10 (VZV) and 0.06 (EBV) (p > 0.1 for each). for 0.1 > (p (CMV), 0.11 (EBV) 0.06 (HSV), and -0.02 (VZV) 0.10 coefficients correlation rank population; There was no significant correlation between HHV6 antibody levels by by levels antibody HHV6 between correlation significant no was There IF and antibody levels to any of the other human herpesviruses in this this in herpesviruses human other the of any to levels antibody and IF r .♦ ...... ♦♦♦♦♦ ...... ♦.♦♦♦ ++♦♦♦ ..... ♦♦♦♦♦ ♦♦♦♦♦ Neg Neg Neg iHHV6 I n5 Ant- F n50) 0 5 in 1 ( IF 6 V H ti-H n A ) 50 in 1 ( IF 6 V H ti-H n A iHHV6 I 1 n5 ) 50 in 1 ( IF 6 V H ti-H n A 4 t ♦ ♦ ♦ ♦ ♦ ♦ 44 ± 4 ++ + 4 ♦ ♦ ♦ r * ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦♦ i ------83 *<5 u > H < w c u i Neg 128 <2 32 64 16 8 4 2 ♦ ♦ ..... ♦♦♦ ♦♦♦+. ♦♦♦♦♦ . . . . . +♦♦♦♦ 44444 44444 44+44 e ± + ± Neg Neg 4 44 4 44 444 444 iHHV6 F n 0 ) 50 in 1 ( IF 6 V H ti-H n A + t ------+ +++ ++ r +++

iue3. 16 3a. Figure Anti-VZV CF titre Anti-HSV CF titre <2 <2 128 competitive RIA and antibody levels to any of the other human herpesviruses herpesviruses human other the of any to levels antibody and RIA competitive by levels antibody HHV6 between correlation significant no was There 0.09 (VZV) and 0.04 (EBV) (p > 0.1 for each). for 0.1> (p (EBV) 0.04 and (VZV) 0.09 (CMV), -0.01 (HSV), -0.13 coefficients correlation rank population; this in and antibody to the other human herpesviruses human other the to antibody and Comparison of anti-HHV6 by competitive RIA competitive by anti-HHV6 of Comparison niHH6 y RIA by HV6 Anti-H niHH6b RIA by HV6 Anti-H

+ O n 84 16 5 32 £ ++ + Neg <2 128128 + + + oooooooo < I ^ »—< * - A nti-H H V 6 by RIA by 6 V H nti-H A A nti-H H V 6 by RIA by 6 V H nti-H A * — n - .—4 o c I-H % inhibition) (% % inhibition) (% ♦ + + ++ ♦ + 4 4

tj 44 444 444 - n i

r T

444 4 1—1 4

+ +

4^4 no 1—1 4 44 4 + +

t—H - r

4 444

& 4 4 + t r 444 4 4 m 00 8 8 i - i r r! r! r i - i 4 _

ON 4 4 iue3. 17 3a. Figure

Anti-VZV CF Titre Anti-HSV CF Titre the to antibody and IF by IgG anti-HHV6 of Comparison <2 32 16 4 2 8 other human herpesviruses in sera from 28 children 28 from sera in herpesviruses human other ♦ ♦ ♦ +♦♦ +♦ ♦ +♦♦ ♦ +♦ ♦ +♦♦ e ± + +♦Ng + + +++ ++ + ± Neg ++♦ ++ + ± Neg Neg 0.23 (CMV, p > 0.1), 0.35 (VZV, p > 0.05) and 0.27 (EBV, p > 0.1). > p (EBV, 0.27 and 0.05) > p 0.1), > (VZV, p 0.35 (HSV, 0.1), > -0.18 p (CMV, 0.23 coefficients correlation rank population; this levels antibody HHV6 between correlation significant no was There by IF and antibody levels to any of the other human herpesviruses in in herpesviruses human other the of any to levels antibody and IF by ♦ ♦ ♦ ♦ ♦♦ ♦ ♦ ♦♦ ♦ ♦ iHHV6 6 V H ti-H n A niHV F( in50 ) Anti-HHV6(1 IF IF ( 1 in in ♦ 50 )

85 o (N Cx2 £Q > U« 2 i 8 U 16 ~ < S > < C C a> + + Neg <2 32 + + 4 . + + ♦ ♦♦ ♦♦ ♦♦+ + ♦♦ + ♦ ♦♦ ♦♦ + ♦ ♦ ♦ ++ + e ± Neg ♦ ♦ ♦ t t + ♦ ♦ ♦ ♦ ♦ + iHHV6 6 V H ti-H n A Anti-HHV6 ♦♦ + + + ♦ + IF IF + IF IF + +++ ++ + ( 1(1 1 * ♦ +♦ in in in + ♦ ♦ ♦ ++ + 50 50) ) ♦ ♦

♦

3a. 7. Detection of anti-HHV6 IgM by indirect IF:

An indirect IF assay for detection of anti-HHV6 IgM was developed in the same way as for the IgG test. None of 30 blood donor sera gave unequivocal staining in this assay. A group of 25 cord blood samples were also tested for HHV6 IgM antibody by IF, with negative results. Childrens' sera were selected for screening for anti-HHV6 IgM by indirect IF since the age of acquisition of anti-HHV6 IgG suggested that most primary HHV6 infections occur during the first year of life. A small group of childrens' sera were identified as anti-HHV6 IgM positive and were used to further refine this assay. All of these HHV6 IgM antibody positive samples were also tested for CMV IgM antibody with negative results. Photomicrographs of indirect IF screening of sera for anti-HHV6 IgM by indirect IF are given in Figure 3a. 18.

3a. 7. i. titres and scoring system

In a similar way to that for the IgG assay, anti-HHV6 IgM positive sera were tested at a fixed dilution (1 in 24) and titrated in doubling dilution (1 in 24 to 1 in 384) to give an end point. All of the childrens' sera initially identified as anti-HHV6 IgM positive had an endpoint titre of 1 in 24 and were designated + in the assay. Sera with a titre of > 1 in 24 were only identified when samples from roseola infantum patients or adults who had reactivated HHV6 were tested for specific IgM antibody (see 3a. 8., 3b. 1. and 3b. 2.). A comparison of titre with the score at 1 in 24 for 15 HHV6 IgM positive and negative serum samples is given in Figure 3a. 19. There were more problems with sera giving non-specific staining in the anti-HHV6 IgM assay than had been experienced when screening samples for HHV6 IgG antibody by indirect IF. Samples designated ± for HHV6 IgM at 1 in 24 could not be resolved by titration but the score for anti-HHV6 IgM positive sera at 1 in 24 was clearly related to the fluorescent end point (rank correlation coefficient 0.98, p < 0.001). A scoring system was adopted for future use and the range of titres for each scoring group are given in Table 3a. 8.

3a. 7. ii. specificity and reproducibility

Five RF positive sera from patients with rheumatoid arthritis and an RF negative serum from a patient who had reactivated HHV6 (patient 21, 3b. 2.) were screened for HHV6 IgM and IgG antibody by indirect IF before and after treatment with RF-absorbans. Results for all sera are given in Table 3a. 9. After absorption all sera were negative for anti-HHV6 IgG and only 1 serum from a rheumatoid arthritis patient contained residual levels of anti-IgG. All sera from rheumatoid arthritis patients were negative for anti- HHV6 IgM before and after treatment with RF-absorbans. Serum from the patient who had reactivated HHV6 in response to a primary CMV infection contained high levels of anti-HHV6 IgG. Virus-specific IgM antibody was only detectable in this sample after

86 Photomicrographs of indirect IF screening of sera for anti-HHV6 IgM

r_. Mm <

++______+++

The fluorescent staining of acetone fixed HHV6 infected J Jhan cells by an anti-HHV6 IgM positive serum is both nuclear and cytoplasmic. Sera were diluted 1 in 24 and cells were counterstained with Evans' blue (x 1600).

87 Figure 3a. 19

A comparison of anti-HHV6 IgM score and titre by indirect IF for 15 sera

+ + + ♦ ♦ ♦ Tf

Z + + ♦ ♦ ♦ 03 L.4> wfj© + ♦♦ ♦ £ (5JD — + ♦♦♦ VO > X X Neg ♦♦♦

t------1------1------1------1------1------r <24 ^24 24 48 96 192 384 >384 HHV6 IgM titre

rank correlation coefficient = 0.98 (p < 0.001)

Table 3a. 8

Range of titres for each group of sera scored for anti-HHV6 IgM by indirect IF

IF score at IF 1 in 50 titre

Neg < 2 4

+ ^ 2 4

+ 24-48

+ + 96-192

+ + + 192->384

88 Table 3a. 9

Anti-HHV6 by indirect IF and RF by latex-agglutination before and after treatment with RF-absorbans

Before absorption After absorption Sem m * Anti-HHV6 Anti-HHV6 * Anti-HHV6 Anti-HHV6 RF IgG score at IgM score at RF IgG score at IgM score at 1 in 50 1 in 24 1 in 50 lin 24

1 + + + + + - + - -

2 + + + + - -- -

3 + + + - -- -

4 + + + - -- -

5 + + --- -

6 - + + + - - - + + +

* + = weak positive, ++ = strong positive for RF

Sera 1 -5 were from patients with rheumatoid arthritis. Serum 6 was from an adult who had reactivated HHV6 (patient 21, see 3b. 2).

89 after treatment with RF-absorbans. These results indicated that false negative IgM reactions due to the presence of high titre HHV6-specific IgG and false positive IgM reactions due to the presence of specific IgG and RF would be minimized by the absorption procedure.

Each batch of slides prepared for the anti-HHV6 IgM assay was tested with reference sera which had been screened and scored previously. While optimizing the indirect IF test for anti-HHV6 IgM, sera were tested in duplicate on at least two separate occasions and the fluorescence was read and scored independently by three different laboratory workers. Once the appropriate scoring system had been established and reproducibility of the system confirmed, samples were tested at a fixed dilution (1 in 24) and read and scored by myself with reference to control sera on each slide. Sera were retested and the score confirmed on a second occasion.

3a. 8. A study of HHV6 antibody in sera from pregnant women and female blood donors

The age of acquisition of HHV6 suggested that the virus is passed from previously infected adults to infants in the first year of life.

A preliminary study of HHV6 antibody in sera from pregnant women was undertaken. The aim of this was to identify any HHV6 primary infection or reactivation in these women which might be relevant to transmission of the virus to their newborn offspring. Serum samples taken at antenatal booking (12-14 weeks gestation) and during the third trimester were available and tested for anti-HHV6 by indirect IF and competitive RIA . Results were compared to those for a group of female blood donors (Table 3a. 10).

The prevalence of specific IgM antibody was found to increase during pregnancy (3 % positive at 12-14 weeks to 31 % positive at 36 weeks, X2 test, p < 0.001). None of a group of female blood donor sera was anti-HHV6 IgM positive by IF. The number of anti-HHV6 IgG positive sera was not significantly different when comparing the female blood donor and pregnant women groups.

The significant levels of anti-HHV6 IgM in sera from pregnant women may be indicative of HHV6 reactivation, because of these results a prospective study of HHV6 in pregnancy was undertaken to investigate this further and to assess the relevance of HHV6 infection to the transmission of the virus (see 3e.).

90 Table 3a. 10

Antibody to HHV6 in sera from pregnant women and female blood donors

Indirect IF (% positive) RIA Sera IgG IgM (% positive) at 1 in 24 at 1 in 50 Female blood donor 0 65 81 (n=43) 12-14 week antenatal 3 51 82 (n=124) 36 week antenatal 31 62 90 (n=32)

The number of anti-HHV6 IgM positives was significantly different when 12-14 week and 36 week sera were compared (4/124 positive compared with 10/32, X2test, P < 0.001).

91 3a. 9. Discussion

The AJ isolate of HHV6 was successfully grown in the J Jhan T-lymphocyte cell line and used as antigen for HHV6 seroprevalence studies. Approximately 60 % of sera from blood donors were anti-HHV6 IgG positive by indirect IF and 80 % were positive by competitive RIA. It was noted that female blood donor sera had a higher prevalence and titre of HHV6 antibody by indirect IF than sera from males, a phenomenon also reported by Clark et al. (1990).

A number of research groups have undertaken HHV6 seroprevalence studies in adults and children using IF or ELISA for detection of specific IgG antibody. Seroprevalence rates quoted for HHV6 in adults have varied greatly. Salahuddin and colleagues (1986) reported that sera from 2 % of healthy adults were HHV6 antibody positive and another group quoted 8 % prevalence of anti-HHV6 IgG in sera from Austrian blood donors (Larcher et al, 1989, Huemer et al., 1989). However, most people now agree that a large proportion of sera from adults is anti-HHV6 IgG positive and that specific antibody levels are generally low in the adult population. Recently quoted seroprevalence rates for adults vary between 35 - 100 % (Tedder et a l, 1987, Brown et al., 1988, Brown et a l, 1988a, Pietroboni et al. 1988, Krueger et al., 1988, Saxinger et al., 1988, Linde et al., 1988, Okuno et al., 1989, Levy et al., 1990a, Clark et al., 1990, Rodier et a l, 1990, Dahlet al., 1990, Enders et al., 1990, Yanagi et al., 1990, Soriano et a l, 1991). Sensitive serological assays are required for the detection of anti-HHV6.

Differences in HHV6 seroprevalence are probably due to problems in the determination of cut-off between positive and negative. There are now many different isolates of HHV6 each of which has its ownin vitro growth characteristics (see 4a. 2). Antigenic differences between these isolates and the time after infection that antigens are prepared may also effect serological results. Where anti-HHV6 ELIS As have been developed (Saxinger et al., 1988, Dahl et al., 1990, Asano et al., 1990) the sensitivity for detection of anti-HHV6 seems to be better than that for IF and is equivalent to that for the competitive RIA.

Serum samples from children were tested for anti-HHV6 IgG by indirect IF in order to establish the age of acquisition of the virus. Maternal antibody was demonstrable in children up to 6 months of age. From age 6 months the prevalence of HHV6 antibody rose until by 1 year of age the HHV6 antibody levels were not significantly different to that found in adults. The early acquisition of HHV6 once maternal antibody has waned has been confirmed by a number of other research groups (Yamanishi et al., 1988, Takahashiet al., 1988, Brown et a l, 1988, Pietroboni et al., 1988, Knowles and

92 Gardner, 1988, Saxinger et al ., 1988, Okuno et al., 1989, Yoshikawa et a l, 1989, Ueda et al.y 1989, Enders et al., 1990, Farr et al, 1990, Asano et al., 1990).

The indirect IF and competitive RIA seem to be specific for the detection of anti-HHV6 in groups of healthy adults and children. There was no correlation between presence or level of HHV6 antibody and presence or level of antibody to any of the other human herpesviruses in these groups. Similar cross-correlation studies have been carried out by other groups, with similar results (Saxinger et al., 1988, Asano et al., 1990). The specificity of serological assays for detection of HHV6 antibody has also been confirmed by cross-absorption of sera from adults and children with different virus preparations (Salahuddin et al., 1986, Saxinger et a l, 1988, Buchbinder et a l, 1989, Okuno e ta l, 1989).

An indirect IF test for the detection of anti-HHV6 IgM was developed. The pre treatment of sera with RF-absorbans removed all trace of anti-HHV6 IgG and also significantly reduced the levels of RF (anti-IgG) in serum samples. This treatment would be expected to minimize false positive and false negative reactions in the anti- HHV6 IgM assay. None of a group of blood donor sera was found to be anti-HHV6 IgM positive by this method but an increase in the prevalence of specific IgM antibody was noted when comparing 12-14 week antenatal sera with those taken late in pregnancy. These results required further investigation and a prospective study of HHV6 in pregnancy was undertaken (3e.). A small number of cord blood samples were tested for anti-HHV6 IgM by indirect IF, all were negative. Farr et al. (1990) tested a large number of cord blood samples for anti-HHV6 IgM, with negative results, and concluded that HHV6 is usually acquired after birth rather than in utero.

Attempts to develop an RIA for the detection of anti-HHV6 IgM proved unsuccessful. Culture of the AJ isolate of HHV6 in the J Jhan, HSB-2 or c8166 cell lines, by the methods described, did not result in the production of supernatant virus. The cell extract used as antigen in the competitive RIA gave very high backgrounds in an IgM capture or indirect immunoassay for anti-HHV6 IgM. Saxinger and co-workers reported use of an indirect ELISA for detection of HHV6 specific IgM antibody (Saxinger et al., 1988, Steeper et al, 1990), but only 6 of 8 IgM positive sera by indirect IF were also positive by ELISA, no further details of sensitivity or specificity of this assay were given. It was hoped that the monoclonal antibodies prepared against the AJ isolate of HHV6 (see 2i. and 3c. 3) would be useful in the development of a specific RIA or ELISA for detection of anti-HHV6 IgM.

93 3b. HHV6 antibody in patient groups

3b. 1. A study of HHV6 antibody in children with roseola infantum

Yamanishi and colleagues (1988) reported isolation of HHV6 and concomitant seroconversion for anti-HHV6 IgG by ACIF in young children with roseola infantum (exanthem subitum), a febrile illness of early childhood characterised by pyrexia and rash.

A study was undertaken to investigate the association of HHV6 infection with roseola infantum and other febrile illnesses in 18 children. Serum samples were taken from 9 infants aged between 2 and 22 months who were examined, and in some cases admitted to hospital (patients 1-3 and 5) because of a recent onset of high fever and rash consistent with a diagnosis of roseola infantum. Patient 10 (a 9 month old child) was described as having pyrexia of unknown origin. The pattern of fever was similar to that found in roseola infantum but no rash was reported. Patient 11 had all the symptoms of roseola infantum but also suffered from gastroenteritis which may have been due to a rotavirus infection. This 5 month old child was hospitalized 2 days after the onset of a head cold when she began to have diarrhoea and vomiting. At this time a rash also began to appear over her cheeks and upper trunk. The child had a temperature of 38.8 °C.

Serum samples from patients 1-11 were tested for anti-HHV6 IgM and IgG by indirect IF at a fixed dilution and, when enough serum was available, for anti-CMV IgM using a commercially available ELISA (Table 3b. 1). Seroconversion for anti-HHV6 IgG was shown in all infants except patient 5 who had specific IgG in the first sample taken. Anti-HHV6 IgM was demonstrated in one or more of the samples tested from each child except patient 4 who had equivocal levels of HHV6 IgM antibody in the first sample. It is possible that the rotavirus gastroenteritis suffered by patient 11 was more severe than would normally be the case because of the concomitant HHV6 primary infection. All serum samples from patients 1-6 were negative for anti-CMV IgM.

Serum samples from 7 cases of childhood pyrexia and rash associated with proven enteroviral or rubella infection (serological confirmation of clinical diagnosis was by other laboratory workers) were also tested for anti-HHV6 IgM and IgG by IF and anti- CMV IgM by ELISA, with negative results (Table 3b. 2). Where serum was available, samples from children with pyrexia and rash (including the roseola infantum cases) were also tested for anti-HHV6 by competitive RIA and the results compared with those for indirect IF. These results are given in Table 3b. 3.

94 Table 3b. 1

Antibody to HHV6 and CMV in sera from 11 children with a clinical diagnosis of roseola infantum

Patient Day serum Anti-HHV6 Anti-HHV6 Anti-CMV (age at onset taken IgG by IgM by IgM by in months) (rash=dayl) indirect IF indirect IF ELISA -2 - - - 1 3 + + - (10) 44 + + + + - 2 -1 + -- (8) 6 + + + + - 3 2 + + - (22) 14 + + + - 4 1 + + - (10) 13 + + - - 5 5 + + + - (2) 22 + + + - 6 1 - + - (10) 29 + + + - 7 1 + - - (17) 30 + + + ND 8 1 - - ND (12) 17 + + - 9 1 -- ND (11) 43 + + + + ND 10 *6 - + ND (9) 76 + + + + ND 11 -2 - - ND (5) 12 + + + + ND

* No rash was reported in this case and day 1= day of onset of pyrexia.

ND = not done

95 Table 3b. 2

Antibody to HHV6 and CMV in sera from 7 children with pyrexia and rash

Patient Day serum Anti-HHV6 Anti-HHV6 Anti-CMV (age at onset taken IgG by IgM by IgM by in months) (rash=dayl) indirect IF indirect IF ELISA 12 3 --- (5) 36 --- 13 2 --- (4) 89 -- - 14 6 + -- (2) 12 + -- 15 2 - + - (5) 9 + -- 16 2 - + - (5) 6 --- 15 + -- 17 31 + -- (2) 87 - - - 18 1 - + - (17) 8 - + -

96 Table 3b. 3

A comparison of anti-HHV6 by indirect IF and competitive RIA in sera from children with pyrexia and rash

Patient Day serum Anti-HHV6 Anti-HHV6 (age at onset taken IgG by by RIA in months) (rash=dayl) indirect IF (% inhibition)

-2 - 16 1 3 + ND (10) 44 + + + 85 4 1 + 25 (10) 13 + + 30 5 5 + + 70 (2) 22 + + 70

8 1 - 25 (12) 17 + ND

9 1 — 20 (11) 43 + + ND

10 *6 — 28 (9) 76 + + 80

11 -2 - 27 (5) 12 + + + 38 14 6 + 80 (2) 12 + 77

15 2 - 28 (5) 9 + ND

16 2 - 37

(5) 5 - 35 15 + 73 17 31 + 60 (2) 87 - 58

18 1 - 26 (7) 8 - ND

* day 1= onset of pyrexia ND= not done

97 Late convalescent samples for patients 1 and 10 (taken 44 and 76 days after onset of illness, respectively) were available and seroconversion for HHV6 antibody by competitive RIA was shown. The second serum sample from patients 4 and 11 was taken much earlier (13 and 12 days after onset of illness, respectively) and seroconversion for HHV6 antibody by competitive RIA could not be demonstrated. It seems from these results that although the competitive RIA is more sensitive for the detection of anti-HHV6 in adults, seroconversion for HHV6 antibody in children can be detected earlier using indirect IF. A possible explanation for these results is that the antibody produced early in infection is of low avidity which is detectable by indirect immunoassays but does not compete efficiently with the high titre radiolabelled anti- HHV6 serum.

Patients 14 and 17 were only 2 months old at the onset of illness and although the IF result was equivocal or negative for anti-HHV6 IgG, all samples from these individuals gave significant inhibition in the competitive RIA. It was assumed these results were due to low levels of maternal antibody in these infants since no HHV6 IgM antibody was detectable and where a late sample was available (patient 17) a drop in the % inhibition was noted. Patient 5 was also only 2 months old at the time of illness and samples taken on day 5 and 22 were both anti-HHV6 positive by IF and RIA, this may also be due to maternal antibody but specific IgM antibody was detected in both samples making a diagnosis of roseola infantum likely.

Patients 1, 8 and 9 were 1 0 -1 2 months old at the time of illness and were assumed to have very little maternal antibody. The first serum from each of these infants was taken before any HHV6 IgM or IgG antibody could be detected by IF. These samples gave 16 %, 25 % and 20 % inhibition of label binding in the competitive RIA, respectively. It was assumed that these levels of inhibition represented those of truely negative serum samples and these results were used to establish the cut-off between positive and negative in the competitive RIA (see 3a. 4. ii. and Table 3a. 4). The first samples taken from patients 11 (age 5 months) and 18 (age 7 months) gave 27 % and 26 % inhibition in the competitive RIA, respectively, again neither HHV6 IgM nor IgG antibody was detectable by IF in these samples.

Yamanishi and colleagues (1988) first reported the association of HHV6 primary infection with roseola infantum. They isolated HHV6 from roseola infantum patients and demonstrated HHV6 IgG antibody seroconversion in these children. These HHV6 isolation and IgG antibody seroconversion results were confirmed by other research workers (Takahashiet a l, 1988, Asano et al ., 1989a, Kikuta et al ., 1989, Yoshiyama et al ., 1990, Enders et a l , 1990, Yoshida et a l , 1990, Portolani et al, 1990, Kido et al, 1990).

98 It was possible, using indirect IF, to demonstrate seroconversion for anti-HHV6 IgG with or without the presence of anti-HHV6 IgM antibody in 11 patients with clinically diagnosed roseola infantum. Four other groups demonstrated HHV6 specific IgM antibody by IF in sera from roseola infantum patients (Knowles and Gardner, 1988, Enders et al ., 1990, Kondo et al., 1990, Irving et al., 1990). One of the patients diagnosed as having roseola infantum (patient 10) had no recognisable rash associated with illness. Suga and coworkers (1989) also reported anti-HHV6 IgG seroconversion (and virus isolation) in an infant with pyrexia but no rash.

Serum samples were available from children with fever and rash associated with enteroviral or rubella infection. None of these samples were positive for anti-HHV6 IgM by indirect IF. Some of the sera from children with pyrexia and rash (including roseola cases) were also tested for anti-CMV IgM, with negative results. Another research group (Enders et al., 1990) also tested the specificity of indirect IF for the detection of anti-HHV6 IgM, they found that children with acute viral infections (including rubella, measles, mumps and varicella) were all negative for anti-HHV6 IgM by IF.

3b. 2. Reactivation of HHV6 in 3 patients

Three cases of HHV6 serological reactivation, one in a 3 year old child and the other two in adults were identified during the course of HHV6 antibody prevalence studies. Multiple samples were available from these patients and were investigated retrospectively.

Patient 19 was a 3 year old child who was reported to have had a throat infection with multiple mouth ulcers which did not respond to penicillin. After this treatment he had a poor appetite and loose pale stools, over the next month his diarrhoea persisted and he failed to gain weight although no microbiological cause for his illness was found. He was admitted to hospital 4 weeks after onset of illness with anterior cervical lymphadenopathy and hepatosplenomegaly. He was reported to have a mononucleosis like illness. A serum sample taken at this time was heterophile antibody positive and a presumptive diagnosis of EBV infection was made. Four months after his illness began he started to improve and put on weight and, in spite of occasional bouts of diarrhoea, he remained healthy. Serum samples taken during his long illness were available to screen for anti-HHV6 IgM and IgG.

Patient 20 was a male aged 50 years who had tender lymphadenopathy, atypical lymphocytosis and persistent pyrexia (39 °C). Blood samples were taken regularly (from 11-167 days after onset) during the investigation of his persisting febrile illness.

99 Patient 21 was a healthy adult found, by random screening, to have a very high anti- HHV6 IgG titre. She had had mild symptoms of neck stiffness and fever (38 °C) 3 months prior to the HHV6 test. Sera were available taken approximately 7 months prior to this episode and from days 1, 42, 71, 92 and 127 after onset of illness.

Sera from patients 19-21 were screened for anti-HHV6 IgM and IgG by indirect IF, the results are given in Table 3b. 4. The score and the titre for HHV6 antibody have been quoted for these individuals so that differences in high titre antibody, that were not demonstrable by testing at a fixed dilution, could be shown. Anti-EBV and anti-CMV serology for these patients was carried out by standard IF and ELISA, respectively by other laboratories and there was not sufficient serum available from these patients to repeat these assays. The first sample taken from patient 19 was taken at the time of admission to hospital, 4 weeks after the child's illness began, and was positive for both anti-HHV6 IgM and IgG. It was positive for anti-CMV IgM and anti-EBV VCA IgM but negative for anti-CMV and anti-EBV VCA IgG. The second sample was taken at about the time the child started to improve and this was again positive for HHV6 IgM and IgG antibody. The patient was now EBV IgM antibody negative but EBV VCA IgG antibody positive, consistent with having had a primary EBV infection. The third sample which was taken 6 months after the illness began was positive for anti-HHV6 and anti-EBV VCA IgG but was negative for IgM antibody to both viruses. The second and third samples from patient 19 were anti-CMV IgG negative. The presence of HHV6 specific IgM antibody in 2 of the 3 samples tested indicates HHV6 primary infection or reactivation during the time span in which illness occurred. As the first sample from this patient was taken so late after the onset of clinical symptoms it was not possible to distinguish between primary infection and reactivation in this case. The patient did not have the 'classical" symptoms of pyrexia and rash and was older than most children who suffer from roseola infantum and so reactivation of HHV6 in response to the primary EBV infection was thought to be most likely in this case. Anti- CMV IgM was detectable in the first sample from patient 19 but, because no anti-CMV IgG was detected in any of the samples, this anti-CMV IgM reaction was likely to have been non-specific.

A serum sample taken from patient 20 on day 11 of his illness contained anti-HHV6 IgG. Anti-HHV6 IgM was detectable on day 18 and there was an eightfold drop in HHV6 IgG antibody between days 25 and 167. Anti-CMV IgM and IgG and anti-EBV VCA IgM were not detected in any sample.

A sample from patient 21 taken 213 days before the onset of illness showed a low level of anti-HHV6 IgG. On the first day of illness low level anti-HHV6 IgM was present and by the time of the second sample after onset (day 42) there had been a 64-fold increase

100 Table 3b. 4

Antibody to HHV6 in 3 cases of HHV6 reactivation

Day serum Anti-HHV6 by IF taken Patient IgG score IgM score (onset of illness IgG titre IgM titre = day 1) at 1 in 50 at 1 in 24 28 640 +++ 48 + 19 167 320 +++ 48 +

237 320 +++ <24 -

11 1,280 +++ <24 - 18 640 +++ 192 ++ 19 640 +++ 192 ++ 20 24 2,560 +++ 96 + 25 2,560 +++ 96 + 167 320 +++ <24 -

-213 160 ++ <24 - 1 160 ++ 48 + 42 10,240 +++ 192 ++ 21 71 10,240 +++ 768 +++ 92 5,120 +++ 192 + 127 5,120 +++

101 in anti-HHV6 IgG together with a 4 fold increase in anti-HHV6 IgM response. IgM antibody to CMV was detected in the first and second samples and seroconversion for anti-CMV IgG was consistent with a primary CMV infection. Coxsackie A9 was isolated from a throat swab taken on day 1 of the patient's illness.

It was possible to show production of HHV6 specific IgM antibody in individuals already seropositive for IgG (patients 20 and 21). The causes of HHV6 reactivation remained unidentified but patient 21 had both a Coxsackie A9 and a primary CMV infection and it is possible that one of these might have been a trigger for reactivation.

Two groups (Kirchesch et al., 1988, Enders et al., 1990) reported concomitant rises in IgM antibody to HHV6 and CMV in patients with acute illness. There have also been reports of CMV infection associated with anti-HHV6 IgG seroconversion (Larcher et al., 1988) and elevated HHV6 IgG antibody titres (Irving et al., 1988, Linde et al., 1988, Morris et al., 1988, Larcher et al., 1989, Chou and Scott, 1990, Dahl et al., 1990, Linde et al., 1990, Irving et al., 1990a, Sutherland et al., 1991). Raised levels of anti- HHV6 IgG with or without the presence of specific IgM antibody have also been demonstrated in some patients with EBV (Niederman et al., 1988, Larcher et al., 1989, Dahl et al., 1990, Enders et al., 1990, Linde et al., 1990) and VZV (Morris et al., 1988) infections. These results probably reflect coinfection and simultaneous reactivation of HHV6 and the other human herpesviruses. There have been reports of a rise in anti- HHV6 IgG with or without detection of specific IgM antibody in adults with no serological evidence of other concomitant infections as in patient 20 (Morris et al., 1988, Niederman et al., 1988, Steeper et al., 1990, Irving and Cunningham, 1990, Irving et al., 1990, Sutherland et al., 1991). Where HHV6 IgM or IgG antibody rises have been associated with simultaneous CMV infection, crossabsorption studies have been undertaken by two groups to check the specificity of anti-HHV6 serological assays. Irving and coworkers (1990a) found no evidence of serological crossreaction between these two viruses but a second group (Sutherland et al., 1991) reported that there were crossreactive antibodies produced in some transplant patients.

3b. 3. HHV6 and chronic fatigue syndrome:

Chronic fatigue syndrome (CFS) has been defined as a clinical entity characterized by chronic debilitating fatigue lasting longer than 6 months (Komaroff, 1988). The term has been used to describe an ill-defined complex of diseases in which fatigue is just one of the symptoms used for clinical diagnosis. In a large proportion of patients, illness starts with acute 'flu-like' symptoms and research has centered on the identification of a chronic viral infection associated with disease. The term post-viral fatigue syndrome (PVFS) has been used by clinicians to describe chronic fatigue in patients with an

102 identifiable 'acute phase" of illness. Myalgic encephalomyelits (ME) has similarities with CFS and PVFS but illness may begin with an 'outbreak" such as that associated with 'Icelandic disease" and 'Royal Free disease "(discussed in Bell et al., 1988). The main symptoms of ME are headache and extreme exhaustion, particularly after exercise, but the secondary symptoms most often reported are different to those described for CFS. In acute 'epidemic" ME brainstem signs such as vertigo, hyperacusis and tinnitus are common (Fegan etaL, 1983, Bell eta l ., 1988).

Until recently EBV or enterovirus infections were thought to be the most likely viral causes of prolonged fatigue (reviewed by Buchwald and Komaroff, 1991). Chronic mononucleosis starts with classical acute IM which then does not resolve and has many similarities with the symptoms of chronic fatigue. Elevated titres to EBV antigens have been described in patients with prolonged fatigue and other symptoms such as fever, myalgia and cervical lymphadenopathy (discussed in Straus, 1988, Komaroff, 1988, Hotchin et al., 1989). Archard and colleagues (1988) demonstrated enterovirus-specific RNA in skeletal muscle biopsies of some patients with PVFS. Two other groups (Yousef et al., 1988, Halpin and Wessely, 1989) were also able to show that up to half of PVFS patients had detectable enterovirus group antigen (VP1) in their circulation. Samples from some of these patients were also tested for EBV antibody. Evidence of active EBV infection was found in 20 % of patients, none of whom were identified as suffering from chronic enterovirus infection (Hotchin et al., 1989). These results suggest that persistent fatigue accompanies a number of viral diseases and that the cause of illness in a group of CFS patients may be multifactorial.

The isolation of HHV6 in 1986 led to renewed interest in the association of lymphotropic herpesvirus infections with symptoms of prolonged fatigue (discussed in Bames, 1986).

3b. 3. i. anti-HHV6 IgG in a group of Paul-Bunnell negative glandular fever patients

A cohort of 40 patients with recurrent glandular fever-like illness and fatigue were screened for anti-HHV6 IgG by indirect IF and anti-CMV, HSV and VZV by CFT. Serology for anti-EBV was carried out by colleagues at Royal Postgraduate Medical School. Patients had suffered recurrent episodes of fatigue, lymphadenopathy, sore throat, myalgia, malaise, arthralgia and depression for several months prior to enrolment in the study but none of them had evidence of recent EBV infection (all were Paul-Bunnell negative). The investigation of potential viral causes for illness in these patients was part of a larger study of symptoms and infections associated with chronic

103 fatigue (Read et al ., 1990). Serum samples were available from the first attendance at the Clinical Research Centre and from the third visit, 3-4 months after referral.

The prevalence of anti-HHV6 IgG in serum specimens from the study group patients and age and sex matched blood donors are summarised in Table 3b. 5. Sera from patients were more likely than controls to be HHV6 antibody positive (74 % compared with 53 %, X2 test, p < 0.05) and even more likely to be strongly positive (60 % compared with 28 %, X2 test, p = 0.01). Most striking was the difference in prevalence of strong positive reactions when comparing sera from male patients and male controls (63% compared with 13%, X2 test, p < 0.01). There was no significant difference in EBV VCA, EBV early antigen (EA), CMV, VZV or HSV antibody when sera from patients were compared with sera from age and sex matched blood donor controls.

Raised antibody titres to a variety of viruses including EBV, HSV, CMV and measles have been described in some groups of CFS patients (discussed in Holmes et al ., 1987, Straus, 1988). In sera from the group of Paul-Bunnell negative glandular fever (PBNGF) patients anti-HHV6 IgG but not antibody to the other human herpesviruses was elevated indicating that these results were not due to non-specific polyclonal activation. There was no significant difference in HHV6 antibody levels when comparing sera taken at referral to samples taken 3-4 months later. This is probably not surprising since the patients had had an identifiable 'acute phase' of illness many months prior to their referral.

The elevated anti-HHV6 in the group of patients described may have been due to a secondary reactivation of the virus which was unrelated to acute illness. It is possible, however that HHV6 had a direct role in the anti-EBV negative glandular fever suffered by these patients.

There have been a number of reports of patients with acute HHV6 infection associated with a mononucleosis-like illness (Krueger et al ., 1987, Niederman et al ., 1988, Steeper et al ., 1990, Irving and Cunningham, 1990). Some of these patients progressed to a chronic disorder with fatigue and recurrent lymphadenopathy, similar to the patient described by Krueger et al. (1987). Kirchesch et al. (1988) also described a patient with glandular fever-like illness who seroconverted to HHV6, but this was associated with a concomitant CMV infection.

There was not sufficient serum from the PBNGF patients to test for anti-HHV6 by the newly developed competitive RIA. In order to try and confirm the association of HHV6 infection with symptoms of prolonged fatigue a second group of patients were examined for evidence of HHV6 infection.

104 Table 3b. 5

Anti-HHV6 IgG by indirect IF in sera from Paul-Bunnell negative glandular fever patients and controls

Patients

II Controls E3 4^ O

All Positive 21 (53%) 31 (74%)

Strongly Positive 11 (28%) 24 (60%)

Female Male Female Male

14/24 7/16 19/24 12/16 All positives (58%) (47%) (79%) (75%)

9/24 2/16 14/24 10/16 Strongly Positive (38%) (13%) (58%) (63%)

Controls are blood donors, age and sex matched with patients.

The number of anti-HHV6 IgG positives by IF was significantly different when sera from patients and controls were compared (74% compared with 53%, X2 test, p < 0.05).

The distribution of high level reactivity (++ and +++) by IF was significantly different when sera from patients and controls (60% compared with 28%, X2 test, p = 0.01) and when sera from male patients and male controls were compared (63% compared with 13%, X2 test, p < 0.01).

105 3b. 3. ii. anti-HHV6 IgG in a group of chronic fatigue syndrome patients

A group of 25 CFS (also described as ME) patients were enrolled into a study group based on the clinical criteria described by Holmes and colleagues (1988). As part of a larger study coordinated by Dr P. Venables (The Kennedy Institute of Rheumatology, London) serum samples from these patients together with age and sex matched controls were tested for anti-HHV6 by indirect IF and competitive RIA. The results are given in Table 3b. 6 and Table 3b. 7.

The group was predominantly female (approximately 85 %) and the % positive for HHV6 antibody was high in both patient and control groups. Neither the prevalence of HHV6 antibody nor the number of anti-HHV6 strong positives by IF or RIA was significantly different when sera from patients and age and sex matched controls were compared. The average % inhibition in the HHV6 RIA for patients and controls together with 95 % confidence limits for the means are given in Figure 3b. 1, the difference between the means of these two groups was not statistically significant.

There was no evidence for an association of HHV6 infection with the symptoms of persistent fatigue suffered by this group of patients. These results contrast with reports of increased prevalence with or without demonstration of a higher mean titre of anti- HHV6 in CFS patients when compared with controls (Komaroff et al., 1988, Straus, 1988, Ablashi et al., 1988, Daugherty etal., 1991, Gupta and Vayuvegula, 1991).

The differences in clinical criteria for enrolment of patients in chronic fatigue study groups first became apparent when Wakefield et al. (1988) reported no difference in anti-HHV6 seroprevalence when comparing ME patients and controls. Another group (Gold et al., 1990) examined patients with chronic fatigue, who had elevated anti-EBV titres, for evidence of concomitant HHV6 infection. They found no difference in HHV6 antibody prevalence and titre when comparing this group with controls. Three other groups also found no difference in HHV6 antibody prevalence and titre when comparing CFS patients with controls (Levy et al., 1990a, Smithet al., 1991, Marshall etal., 1991).

3b. 4. HHV6 in AIDS patients

It has been suggested that cofactors may affect the prognosis of human immunodeficiency virus (HIV) infection and the interval between HIV seroconversion and the onset of clinical disease. Life-threatening opportunistic infections and neoplasms occur frequently in patients with acquired immune deficiency syndrome (AIDS). The interaction of HIV with other infectious agents may explain why some

106 Table 3b. 6

Anti-HHV6 by indirect IF in sera from chronic fatigue syndrome patients and controls

n = 50 Neg + + + + + + + Strong Positive

3 2 7 8 5 Patients n = 25 (12%) (8%) (28%) (32%) (20%)

2 5 8 6 4 Controls n = 25 (8%) (20%) (32%) (24%) (16%)

Controls are healthy adults, age and sex matched with patients.

Patients: Anti-HHV6 IgG positive sera = 80% Anti-HHV6 IgG strong positive sera = 52%

Controls: Anti-HHV6 IgG positive sera = 72% Anti-HHV6 IgG strong positive sera = 40%

There was no significant difference in the number of anti-HHV6 IgG positives by IF when sera from patients and controls were compared (80% compared with 72%, X2test, p > 0.1).

There was no significant difference in the distribution of anti-HHV6 IgG high level reactivity (++ or +++) by IF when sera from patients and controls were compared (52% compared with 40%, X test, p > 0.1).

107 Table 3b. 7

Anti-HHV6 by competitive RIA in sera from chronic fatigue syndrome patients and controls

% inhibition and score n = 50 0-30 31-40 41-60 61-80 1 81-100 Neg ± + + + +++ Strong Positive

3 2 5 14 1 Patients n = 25 (12%) (8% ) (20%) (56%) (4%)

1 3 5 13 3 Controls n = 25 (4% ) (12%) (20%) (52%) (12%)

Controls are healthy adults, age and sex matched with patients.

Patients: Anti-HHV6 positive sera = 80% Anti-HHV6 strong positive sera = 60%

Controls: Anti-HHV6 positive sera =84% Anti-HHV6 strong positive sera = 64%

There was no significant difference in the number of anti-HHV6 positives by RIA when sera from patients and controls were compared (80% compared with 84%, X2test, p > 0.1).

There was no significant difference in the distribution of anti-HHV6 high level reactivity (++ or +++, > 60% inhibition) by RIA when sera from patients and controls were compared (60% compared with 64%, X2test, p > 0.1).

108 iue3. 1 3b. Figure There was no significant difference in the m ean % inhibition inhibition % ean m the in difference significant no was There for each group. each for when sera from patients and controls were compared compared were controls and patients from sera when patients. CFS with matched sex and age adults, healthy are Controls (62% compared with 60%, t-test, p > 0.1). > p t-test, 60%, with compared (62% ) —| x |— ( given are ean m the for limits confidence and Mean % inhibition % 100 20 30- 40- 70- 80 - 50- 60- 90 - 10 inhibition in the anti-HHV6 competitive RIA competitive anti-HHV6 the in inhibition o- - - - for sera from chronic fatigue syndrome fatigue chronic from sera for patients and controls and patients ains controls patients n = 25 n = 25 = n 25 = n 109 HIV infected individuals progress to AIDS soon after seroconversion while others remain well for long periods of time.

Herpesviruses have been implicated as cofactors for the development of HIV related disease. Studies have shown that they are capable of transactivating HIV in vitro (Rando et al., 1987, Pagano et al., 1988) and coinfection of lymphocytes with HIV and CMV (Skolnik et ah, 1988) and HIV and EBV (Pagano et al., 1988) have been demonstrated. Immunocompromised patients suffer from more severe primary or recurrent herpesvirus infections than healthy individuals and there is some evidence for the in vivo exacerbation of HIV during these infections. A number of groups have reported CMV infection associated with a higher prevalence of p24 antigenaemia and a more rapid progression to AIDS and death in HIV infected individuals (Quinnan et al, 1984, Fiala et al., 1986, Webster et al., 1989).

Many of the first isolates of HHV6 were obtained by culture of peripheral blood mononuclear cells (PBMC) from AIDS patients (Salahuddin et al., 1986, Downing et al, 1987, Tedder et al., 1987) and it was later established that the virus shows tropism for CD4 positive T-lymphocytes in vitro (Lusso et al., 1987, Ablashi et al., 1987). One research group postulated that HHV6 was a cofactor for the development of HIV related disease (Krueger and Ablashi, 1987, Ablashi et al., 1988a). They reported a higher prevalence and mean titre of HHV6 antibody in sera from HIV infected homosexual men than in sera from HIV uninfected controls.

3b. 4. i. anti-HHV6 IgG in sera from HIV-I infected and uninfected homosexual men

In order to try and find evidence for the potentiation of HIV infection by HHV6, sera from HIV-1 infected and uninfected homosexual men were tested for anti-HHV6 IgG by indirect IF and the results compared with antibody levels found previously in male blood donors. Single sera from 90 patients routinely referred to the diagnostic HTV service were tested for HHV6 antibody: 30 sera were anti-HIV-1 positive and contained p24 antigen, 30 were HIV-1 antibody positive but antigen negative, and 30 were negative for both anti-HIV-1 and p24 antigen. No patient was on zidovudine at the time the sera were taken. Anti-HHV6 IgG results for these groups are given in Table 3b. 8.

No significant difference in either HHV6 antibody prevalence or titre was found when the three different groups of homosexual men were compared and the antibody profile was similar to that found previously in male blood donors of a similar age. Although p24 antigenaemia is not an exact measure of HIV replication, if HHV6 did potentiate

110 Table 3b. 8

Anti-HHV6 IgG by indirect IF in sera from HIV-1 infected and uninfected homosexual men

HIV (p24) HHV6 IgG score at 1 in 50 status of ++ study group or + + Neg + + +

Antigen -ve 5 9 8 8 Antibody -ve (17%) (30%) (27%) (27%) CO o G II

Antigen -ve 4 9 11 6 Antibody +ve (13%) (30%) (37%) (20%) n = 30

Antigen +ve 4 9 11 6 Antibody +ve (13%) (30%) (37%) (20%) n = 30

All Groups 13 27 30 20

(14%) (30%) (33%) (22%) n = 90

Male Blood 6 18 13 15 Donors (12%) (35%) (25%) (29%) n = 52

There was no significant difference in HHV6 antibody prevalence or titre when any of the groups were compared (X2 test, p > 0.1).

I ll HIV infection a higher prevalence and perhaps titre of HHV6 antibody in the p24 antigen positive than in the antigen negative groups would have been expected.

Brown et al. (1988a) confirmed these results in a larger study of HHV6 antibody in HIV-1 antibody positive and negative sera from blood donors. They reported an increased prevalence and titre of CMV but not HHV6 antibody when comparing the anti-HTV positive and negative groups. A number of other independent serological studies were then published which concluded that HHV6 infection was not a risk factor for HTV-1 seroconversion or progression to AIDS (Holmberg et al. , 1989, Levy et al ., 1990a, Rodier et al, 1990, Enders et al ., 1990, Ceccherini-Nelli et al ., 1990, Soriano et al., 1991). Three research groups published reports in which there was a significantly increased HHV6 antibody prevalence with or without a demonstrable increase in mean titre of HHV6 antibody when comparing sera from HIV infected and uninfected patients (Ablashi et al., 1988, Ablashi et al., 1988a, Krueger et al., 1988a, Huemer et al., 1989, Scott and Constantine, 1990). In each of these reports the HHV6 antibody prevalence and average titre for control sera was generally lower than that reported by other research groups. This may indicate differences in the interpretation of a positive result by IF.

3b. 4. ii. anti-HHV6 in sera from HIV-1 infected and uninfected haemophiliacs

A group of 10 HIV-1 antibody positive sera (patient mean age = 30, range 18-49 years) and one of 10 HIV-1 antibody negative sera (patient mean age =12, range 5-27 years) from haemophiliacs were tested for anti-HHV6 by indirect IF and competitive RIA. Although neither the p24 antigen nor the clinical status of the HIV-1 infected haemophiliacs was known all patients had seroconverted for anti-HIV-1 more than 5 years prior to the test serum being taken. The HHV6 antibody results for these patients are given in Table 3b. 9 and Table 3b. 10.

There was no serological evidence for the potentiation of HIV infection by HHV6 in haemophiliacs. The number of strong positives (++ or +++) for anti-HHV6 was higher in sera from the HIV uninfected than the infected haemophiliacs. By IF, this difference was not significant (6 / 10 compared with 2 / 10, Fisher's exact test p = 0.075). By RIA, 9/10 HIV negative samples were strongly positive for anti-HHV6 (> 60 % inhibition) compared with 1/10 HIV antibody positive samples (Fisher's exact test, p < 0.001). The average % inhibition in the anti-HHV6 RIA for the 2 groups together with 95 % confidence limits for the means are given in Figure 3b. 2. There was a highly significant difference in anti-HHV6 by RIA when sera from HIV-1 infected and uninfected haemophiliacs were compared (51 % compared with 71 %; t-test, p < 0.001).

112 Table 3b. 9

Anti-HHV6 IgG by indirect IF in sera from HIV-1 infected and uninfected haemophiliacs

n = 20 Neg + + + + + + + Strong Positive

HIV-1 1 3 4 2 0 infected n = 10 (10%) (30%) (40%) (20%) (0%)

HIV-1 0 2 2 6 0 uninfected (20%) (20%) (60%) (0%) n = 10 (0%)

HIV-1 infected haemophiliacs: Anti-HHV6 IgG positive sera 6/10 = 60% Anti-HHV6 IgG strong positive sera 2/10 = 20%

HIV-1 uninfected haemophiliacs: Anti-HHV6 IgG positive sera 8/10 = 80% Anti-HHV6 IgG strong positive sera 6/10 = 60%

There was no significant difference in the distribution of anti-HHV6 IgG high level reactivity (++ or +++) by IF when sera from HIV-1 infected and uninfected haemophiliacs were compared (20% compared with 60%, Fisher's exact test, p > 0.05).

113 Table 3b. 10

Anti-HHV6 by competitive RIA in sera from HIV-1 infected and uninfected haemophiliacs

% inhibition and score n = 20 0-30 31-40 41-60 61-80 1 81-100 Neg + + ++ 1 +++ Strong Positive HIV-1 0 1 8 1 0 infected o s II (0%) (10%) (80%) (10%) (0%)

HIV-1 0 0 1 9 0 uninfected n = 10 (0%) (0%) (10%) (90%) (0%)

HIV-1 infected haemophiliacs: Anti-HHV6 positive sera 9/10 = 90% Anti-HHV6 strong positive sera 1/10 = 10%

HIV-1 uninfected haemophiliacs: Anti-HHV6 positive sera 10/10 = 100% Anti-HHV6 strong positive sera 9/10 = 90%

There was a significant difference in the distribution of anti-HHV6 high level reactivity (++ or +++, > 60% inhibition) by RIA when sera from HIV-1 infected and uninfected haemophiliacs were compared (10% compared with 90%, Fisher's exact test, p < 0.001).

114 Figure 3b. 2

% inhibition in the anti-HHV6 competitive RIA for sera from HIV-1 infected and uninfected haemophiliacs

100 -

9 0 -

8 0 -

7 0 - G O • ^ 60 - £ 5 0 -

4 0 -

3 0 -

2 0 -

10 -

o -

HIV-l HIV-1 uninfected infected

Mean and confidence limits for the mean are given ( |— x—| ) for each group.

There was a significant difference in the mean % inhibition when sera from HIV-1 infected and uninfected haemophiliacs were compared (51% compared with 71%, t-test, p < 0.001).

115 The lower prevalence and titre of HHV6 antibody in sera from HIV-1 infected than in uninfected haemophiliacs was a surprising result. It is possible that the age differences between the 2 patient groups introduced a bias into the results but this was thought to be unlikely. When 252 children aged 1-17 years were tested for anti-HHV6 IgG by indirect IF 55 % were antibody positive and 22 % were scored ++ or +++ and designated strongly positive. These figures are not significantly different to those for the group of 96 blood donors (59 % positive, 28 % strong positive).

Interestingly, three other groups have also reported a lower prevalence and or titre of anti-HHV6 in sera from HIV infected patients when compared with controls (Brown, et al., 1988, Spira, et al., 1990, Ceccherini-Nelli et al., 1990). The later group reported a significantly higher prevalence of anti-HHV6 IgG in anti-HIV-1 negative than in anti- HIV-1 positive sera from haemophiliacs. No such difference was noted when HIV-1 antibody positive and negative sera from intravenous drug abusers were screened for anti-HHV6 IgG. Holmberg and colleagues (1989) obtained sequential serum samples from homosexual men, half seroconverted for anti-HIV and the others remained HIV antibody negative. There was no difference in anti-HHV6 prevalence or titre when sera from these two groups were compared but HIV-1 antibody positive / HHV6 antibody positive homosexual men were significantly less likely to develop AIDS than HIV-1 antibody positive / HHV6 antibody negative homosexual men. In the study by Spira et al. (1990) the HIV-1 infected homosexual or bisexual men were divided into those with lymphadenopathy and those with AIDS. There was no significant difference in HHV6 antibody prevalence when sera from the patients with lymphadenopathy was compared with sera from the HIV antibody negative group. The HHV6 antibody prevalence in sera from the AIDS patients was significantly lower than that in the other two groups combined.

These results may partially explain the conflicting reports of HHV6 antibody in HIV infected patients, in many instances the clinical status of the patients was not quoted. It is possible that a reduction in HHV6 antibody prevalence and titre is only demonstrable in sera from patients with AIDS. This phenomenon may be related to the greatly depleted CD4 positive cells in these individuals resulting in reduced expression of HHV6 antigens and a drop in specific antibody levels.

3b. 5. Discussion

Anti-HHV6 serological assays have been used extensively in the identification of clinical illness associated with HHV6 infection in children and adults. The association of HHV6 primary infection with roseola infantum in young children was first reported by Yamanishi and colleagues (1988). It was possible to confirm these results by the

116 demonstration of anti-HHV6 IgM and IgG seroconversion in samples from children with clinically diagnosed roseola infantum. Sera from young children with rash and fever associated with enteroviral or rubella infection were negative for anti-HHV6 IgM and IgG confirming the specificity of these anti-HHV6 serological assays. The demonstration of anti-HHV6 IgM could be useful in the confirmation of clinically diagnosed roseola infantum.

Serological evidence for primary or reactivated HHV6 infection in a child with an EBV-associated mononucleosis-like illness and diarrhoea was presented (patient 19). It was possible to show the production of IgM anti-HHV6 in sera from adults already anti- HHV6 IgG positive. One of these adults (patient 21) had a concomitant primary CMV infection at the time of HHV6 serological reactivation. The detection of anti-HHV6 IgM and a rise in anti-HHV6 IgG in sera from the other adult (patient 20) was not associated with any other demonstrable virus infection. It was concluded from these results that a rise in anti-HHV6 IgG titre and the appearance of HHV6 specific IgM antibody in older children and adults must be interpreted with caution when seeking an aetiological association for HHV6 in patients with illness. The possibility of a secondary reactivation of HHV6, unrelated to primary illness must be considered.

Two groups of patients with prolonged fatigue were examined for serological evidence of recent HHV6 infection. Sera from PBNGF patients were found to have a higher prevalence and titre of HHV6 antibody by indirect IF than sera from age and sex matched controls. Antibodies to the other human herpesviruses were not elevated and there was no evidence of any other acute infection in these patients. It would seem from these data and that of other research workers that HHV6 infection is associated with a mononucleosis-like illness in some adults with or without prolonged symptoms of fatigue. A primary role for HHV6 in the pathogenesis of disease in these patients has not, however been formally proved.

There was no significant difference in HHV6 antibody prevalence or titre, by IF or competitive RIA, when sera from a group of CFS patients were compared with age and sex matched control sera. Different clinical criteria for enrolment of patients into the PBNGF and CFS study groups may account for the discrepant anti-HHV6 results when comparing these two groups of patients. These results do not preclude association of HHV6 with some cases of chronic fatigue. Enterovirus and EBV infections have both been implicated in cases of chronic fatigue and there are probably many other causes of similar symptoms. Patients enrolled into study groups on the basis of similar symptoms of persistent fatigue may have different underlying causes for their illness.

117 Serum samples from HIV-1 infected and uninfected individuals were tested for anti- HHV6. There was no serological evidence for the potentiation of HIV-1 infection by HHV6 in either haemophiliacs or homosexual men. The average % inhibition for anti- HHV6 by competitive RIA was significantly lower in HIV-1 antibody positive than in HIV-1 antibody negative sera from haemophiliacs. The low level of HHV6 antibody in sera from some HIV infected patients may be related to their clinical status.

118 3c. Detection of HHV6 DNA and protein in healthy individuals

Methods for the detection of virus-specific DNA and protein are important for identification of infected tissues and body fluids and for the localization of virus within infected cells. Nested PCR and in situ hybridization techniques were developed for the detection of HHV6-specific DNA. Monoclonal antibodies were prepared against tissue culture derived HHV6 and used for immunohistochemical staining of HHV6 proteins.

3c. 1. Nested PCR for detection of HHV6 DNA:

The polymerase chain reaction (PCR) is a technique for the primer directed enzymatic amplification of specific DNA sequences. Two oligonucleotide primers, the sequences of which flank that of the target DNA, are used for PCR amplification. These primers hybridize to opposite strands of the DNA and are orientated so that DNA synthesis proceeds across the region between the primers. Repeated cycles of heat denaturation of the target DNA, annealing of primers to their complementary sequences and extension of the annealed primers with DNA polymerase are carried out. The amount of specific target DNA theoretically doubles for each cycle since the primers hybridize to the newly synthesized extension products as well as to the original target DNA. One of the first applications of this amplification technique was in the analysis of genetic diseases such as 6-thalassemia and sickle cell anaemia (Saiki et al ., 1985, Saiki et al ., 1986, Mullis and Faloona, 1987).

Originally the Klenow fragment of Escherichia coli DNA polymerase 1 was used in the PCR reaction but because of the heat denaturation step fresh enzyme had to be added for each PCR cycle. Once the thermostable DNA polymerase from Thermus aquaticus (Taq polymerase) was isolated and made available for use in PCR the technique became much more widely used. As well as avoiding the necessity for addition of fresh enzyme for each cycle the use of Taq polymerase (optimum temperature approximately 75 °C) allowed higher temperatures to be used for the annealing step with a subsequent increase in the specificity of the reaction (Saiki et al., 1988).

Since the advent of automated thermal cyclers, PCR has been used extensively for the sensitive detection of virus-specific nucleic acid, particularly in cases where only small amounts of target DNA are available. Amplification of DNA by PCR has been used for confirmation of herpesvirus infections without the need for virus isolation. The technique has been used for the early diagnosis of herpes simplex encephalitis in cerebrospinal fluid (CSF) samples (Powell et al., 1990, Rowley et al., 1990, Aurelius et al., 1991) and the identification of asymptomatic herpes simplex genital infection of

119 pregnant women (Hardy et al., 1990). Cytomegalovirus specific sequences have been amplified from the urine of newborn babies (Demmler et al, 1988) and renal transplant recipients (Olive et al., 1989).

Nested PCR is an adaptation of the original PCR method and was first described by Mullis and Faloona (1987). The method utilizes two pairs of specific nucleotide primers for the amplification of target DNA. The first round PCR is carried out with the outer set of primers and then a portion of the product is transferred to a fresh PCR mixture containing the internal primer pair. The use of two pairs of primers, each of which is specific for the required sequence, enables the analysis of PCR products by agarose gel electrophoresis without the need for confirmatory tests such as Southern blotting. Nested PCR has been used for the detection of viral sequences with enhanced sensitivity over a single round PCR reaction (Garson et a l, 1990, Porter-Jordan et al., 1990, Jarrett et al., 1990, Aurelius et al., 1991, Simmonds et al., 1991, Brytting et al., 1991).

Published viral sequences and PCR data were used for the development of nested PCR for the detection of HHV6-specific DNA.

3c. 1. i. selection of primer sequences

Sequence information for part of the HHV6 genome was available (Lawrence et al., 1990) and was used for the design of specific oligonucleotide primers corresponding to the 13R gene. This area of the genome was selected as a region which was conserved amongst the few HHV6 isolates that were available at this time (Dr R. Honess, personal communication). The 13R primers (H6 and H7) were shown to amplify selectively HHV6 DNA when a single round PCR reaction and Southern blot of the PCR products was carried out (Gopal et al., 1990). Primer sequences are given in Table 2. 1.

3c. 1. ii. optimization of reaction conditions

Initial reaction conditions used for PCR were as described by Saiki (1989). A plasmid (PHD5) containing the 13R HHV6 sequence was provided by Dr R. Honess and used to assess the sensitivity of HHV6 amplification using the 13R primers. The copy number of the plasmid was estimated from the optical density and a ten-fold dilution series was made (50,000 to 0.5 copies) in HHV6 negative DNA. Anneal temperature, primer concentration, magnesium concentration and number of cycles were all adjusted to find the optimum reaction conditions for first and second round HHV6 PCR.

An anneal temperature of 60 °C was found to give optimal results for both first and second round PCR reactions. There were no non-specific bands when the PCR products

120 were analyzed by agarose gel electrophoresis and all second round products gave bands of maximum strength. When the first round anneal temperature was reduced to 55 °C some secondary bands were visible without any enhancement in the detection of HHV6-specific DNA. Increasing the anneal temperature of the first round reaction to 65 °C reduced the sensitivity of the reaction by approximately 10 fold. Adjusting the second round anneal temperature to 55 °C did not alter the results.

Initially, 150 ng of the H6 and H7 primers were used in a reaction volume of 50 |il for the first round PCR. Using this primer concentration 50 copies of HHV6 DNA were detectable after one round of PCR and the 500 copy dilution of plasmid gave a maximum strength band when analyzed by agarose gel electrophoresis. When the primer concentration was reduced to 75 - 100 ng the same results were obtained and 100 ng of the H6 and H7 primers (approximately 13 pmoles) were used in subsequent PCR reactions. Higher concentrations of primers (> 150 ng of each) were found to be inhibitory in the first round PCR reaction with a 10 fold loss in sensitivity and a weak 500 copy band. A gel showing the effect of primer concentration on the strength of the 500 copy band after first round PCR is given in Figure 3c. 1 A.

In the same way as for the first round primers, the concentration of nH6 and nH7 was adjusted between 75 and 250 ng / 50 pi. The optimum concentration of second round primers (nH6 and nH7) was 200 ng (approximately 26 pmoles of each) / PCR reaction. This concentration of primers gave a maximum strength band for all dilutions of plasmid (50,000 to 5 copies).

The magnesium concentration of the PCR reaction buffer was adjusted between 0.75 mM and 2.50 mM for both first and second round PCR reactions. When 1.25 mM or less of magnesium chloride was used both first and second round band strengths were reduced and a 10 - 100 fold reduction in sensitivity was noted. When > 1.75 mM magnesium was used there was an increase in the strength of primer dimer bands without a loss in sensitivity. Results for 1.50 - 1.75 mM magnesium were indistinguishable. Saiki (1989) reported that excess magnesium generally results in the accumulation of non-specific reaction products and so 1.50 mM magnesium chloride was subsequently used in all PCR reactions. A gel showing the effect of magnesium concentration on the strength of the 500 copy plasmid band after first round PCR is given in Figure 3c. IB.

Thirty-five cycles of PCR gave maximum sensitivity for the detection of first round HHV6 products, increasing the number of cycles beyond 35 did not alter the results significantly. Twenty-five cycles of PCR was sufficient for detection of second round products by agarose gel electrophoresis with maximum strength bands.

121 Figure 3c. 1 The effect of primer and magnesium concentration on the amount of amplified PCR product 1 2 3 4 5 6 A

First round PCR products amplified from 5(X) copies of HHV6 plasmid and analyzed by agarose gel electrophoresis. H = HHV6 specific product (223 bp), lane 1 = 0X174 / Hae III molecular weight marker.

A: lane 2 = 200 ng, lane 3 = 175 ng, lane 4 = 150 ng, lane 5 = 100 ng, lane 6 = 75 ng each of H6 and H7 primers. All other conditions unchanged (1.5 mM magnesium). P = primer band.

B: lane 2 = 1.75 mM, lane 3 = 1.50 mM, lane 4 = 1.25 mM, lane 5 = 0.75 mM magnesium. All other conditions were unchanged (1 (X) ng of each primer). The concentration of nucleotides (200 pM of each) used in the first and second round PCR reactions were as described by Saiki (1989). One unit of Taq polymerase was found to be sufficient for the amplification of the HHV6-specific first and second round products in a volume of 50 pi.

Using the optimum conditions for nested PCR, 50 copies of HHV6 13R sequence were detectable after the first round and 5 copies after the second round reaction. The first round PCR was semi-quantitative and the strength of the band was related to the number of copies of HHV6 DNA. First round PCR products were scored ±, +, ++ or +++ depending on the strength of the band when analyzed by agarose gel electrophoresis. All second round bands where of maximum strength. First and second round HHV6 PCR products analyzed by agarose gel electrophoresis are shown in Figure 3c. 2.

DNA extracted from HHV6 infected J Jhan cells was titrated out and amplified by nested PCR. The dilution of DNA which gave a weak first round PCR band and a maximum second round band when analyzed by agarose gel electrophoresis was used as positive control in subsequent PCR reactions. This dilution of DNA gave similar results to the 50 copy dilution of HHV6 plasmid when analysed by agarose gel electrophoresis. A weak positive control was used to ensure that the sensitivity of the nested PCR reaction was consistent.

3c. 2. In situ hybridization for detection of HHV6 DNA:

In situ hybridization is a method for the detection and localization of nucleic acid sequences within a tissue or cell sample. A labelled DNA or RNA probe specific for the required sequence is allowed to anneal to the target nucleic acid. Any labelled probe which has hybridized with the nucleic acid sequence of interest may then be visualized using autoradiographic or enzymatic techniques.

In many situations the detection of virus-specific sequences is not enough to confirm an association of infection with disease. In situ hybridization has some advantages over methods such as Southern blotting, northern blotting, dot blotting and PCR which require the extraction of total nucleic acids from tissues or cells before hybridization. These methods are sensitive for the detection of low copy numbers of viral nucleic acid within a sample but do not give an indication of the number of infected cells. Occasional infected cells may be Tost' in a nucleic acid extraction but are easily recognisable against a background of uninfected cells by in situ hybridization.

Initially, radiolabelled probes were used for in situ hybridization but these have largely been superceded by non-isotopic labelled nucleic acids. Biotin labelled polynucleotides

123 Figure 3c. 2

Nested PCR amplification of HHV6 plasmid using optimal reaction conditions

1 2 3 4 5 6 7 A

First (A) and second (B) round PCR products amplified from HHV6 plasmid and analyzed by agarose gel electrophoresis; lane 1 = d>X174 / Hae III molecular weight marker, lane 2 = negative control (0.5 pg J Jhan DNA), lane 3 = 5,000, lane 4 = 500, lane 5 = 50, lane 6 = 5. lane 7 = 0.5 copies of HHV6 plasmid diluted in 0.5 pg J Jhan DNA.

H = HHV6 specific product (223 bp first round, 173 bp second round).

124 were first described by Langer and colleagues (1981) and have been used successfully for the fine mapping of chromosomal genes with equivalent sensitivity to autoradiographic techniques (Garson et al., 1987). Similar methods have been applied to the detection of virus-specific sequences in cells and paraffin embedded tissue sections (Brigati et al., 1983, Loning et al., 1986, Bums et al., 1986, Lewis et al., 1987, Coates etal., 1987, Anderson etal., 1988).

A method for the in situ localization of HHV6-specific DNA in tissue culture cells and tissue sections was developed by modification of the published methods for viral detection using biotinylated probes.

3c. 2. i. reproducibility of the biotinylation reaction

The PSMD14 plasmid containing 14 kb of HHV6 DNA was biotinylated by nick translation for the in situ localization of HHV6-specific DNA. The PUC8 plasmid without a viral insert was also biotinylated for use as a negative control probe. In order to assess the reproducibility of the labelling reaction and to check that the incorporation of biotin was similar for virus and control plasmids each batch of labelled probes was diluted, spotted onto nitrocellulose and detected as described previously (2h. 2.). Less than 5 pg of biotinylated DNA was routinely detectable by this method. An example of a nitrocellulose filter showing dilutions of probe DNA is given in Figure 3c. 3.

3c. 2. ii. detection of the biotinylated probe

The method for the in situ localization of HHV6 DNA within infected tissues was developed using infected and uninfected J Jhan cells embedded in gelatin and wax and sectioned. The biotinylated HHV6 probe reacted strongly with sections of HHV6 infected J Jhan cells but not with uninfected cells. The specific staining was discrete in the nucleus and more diffuse in the cytoplasm.

In order to confirm that the in situ hybridization method developed was suitable for the detection of viral DNA in tissue sections a commercially available biotinylated probe was used for the detection of CMV DNA in lung tissue from an AIDS patient with CMV pneumonitis. Positive staining was seen in both the nuclear inclusions and cytoplasm of cells with classical 'owl's eye" inclusions. Some smaller CMV infected cells showed nuclear staining only. No staining of these infected cells was seen when using the HHV6 or control plasmid biotinylated probes. The biotinylated CMV and control plasmid probes did not react with HHV6 infected or uninfected J Jhan cells. Initially NBT / BCIP was used as substrate for the in situ hybridization reaction. As this substrate is blue/brown the sections could not be effectively counterstained with haematoxylin. To visualize unstained cells the sections were viewed using phase

125 Figure 3c. 3

Titration and detection of biotinylated probes on a nitrocellulose filter

PSMD14 PUC 8 1 6 1 6 * 0 » 0

2 7 2 7

# 0 • : 0 3 8 3 8 • • 4 9 4 9 • 0 5 10 5 10

• 0

PSMDI4 = HHV6 plasmid. PUC8 = negative control plasmid. I = 10 ng. 2 = I ng. 3 = 500 pg. 4 = 100 pg. 5 = 50 pg, 6 = 20 pg, 7 = 10 pg, 8 = 5 pg, 9 = 2.5 pg plasmid DNA. respectively. 10 = negative control (no plasmid).

126 contrast microscopy with a blue filter. Later, hexazotized new fuchsin and napthol AS- TR phosphate (new fuchsin), a red substrate, was found to detect biotinylated probe with equivalent sensitivity to NBT / BCIP. The staining pattern for HHV6 DNA in control cells and tissues with NBT / BCIP was similar to that for new fuchsin. Photomicrographs of in situ localization of HHV6 DNA in control tissues using new fuchsin as substrate are given in Figure 3d. 1.

3c. 3. Immunohistochemical staining for detection of HHV6 protein:

Immunohistochemical staining is a method for the localization of a specific protein within cell or tissue preparations. Either a polyclonal or monoclonal antibody directed against the protein of interest is required for this technique. Any primary antibody which has bound to the target protein is detected using an anti-species immunoglobulin and this is visualized using enzymatic techniques. Immunohistochemical staining has been used for the detection of viral antigens in infected tissue culture cells and tissue sections. In some cases dual staining for virus-specific nucleic acid by in situ hybridization and virus-specific protein by immunohistochemical staining has been carried out (Wolber and Lloyd, 1988, Porter et al ., 1988). Where MAbs against immediate early (a) herpesvirus proteins are available immunohistochemistry may be used for the rapid detection of CMV in cell culture (Swenson and Kaplan, 1985) and for the identification of latently infected tissues (Toorkey and Carrigan, 1989). The identification of active herpesvirus infections requires the use of antibodies directed against late (y) proteins.

A panel of anti-HHV6 MAbs was prepared and used for immunohistochemical staining of HHV6 specific protein.

3c. 3. i. hybridomas secreting anti-HHV6

Kohler and Milstein (1975) first described a technique for the production of 'immortal" clones of cells secreting a single antibody of interest by fusing normal antibody producing cells with a B-cell tumour line. This technique has been invaluable for the production of virus-specific reagents for detection of antigen and for the development of serological assays.

Mice were immunized with an antigen prepared from HHV6 infected J Jhan cells and the spleen cells fused with a mouse myeloma cell line (Kearney et al ., 1979). This fusion reaction was very successful with visible hybridoma colonies in 85 % of wells after 2 weeks. Twenty-one wells containing hybridoma cells secreting anti-HHV6 were identified by indirect IF, 16 of these antibodies were of the IgG class (14 IgGi and 2 IgG2a) and 5 were of the IgM class. Hybridoma cells from these parent wells were

127 further cloned to ensure that each hybridoma colony was derived from a single cell. An average of 3 clones from each parent well was selected and expanded further. Three different patterns for anti-HHV6 monoclonal staining by indirect IF are shown in Figure 3c. 4.

3c. 3. ii. detection of the primary monoclonal antibody

The immunohistochemical staining method for HHV6 protein in infected tissues was developed using sections of infected and uninfected J Jhan cells in a similar way as for in situ hybridization. Patterns for MAb staining of infected cells by immunohistochemistry were similar to those seen by indirect IF. Some of the staining was very discrete and difficult to visualize except under high magnification. For this reason a pool of 5 HHV6 MAbs (each giving a different staining pattern on tissue culture cells) was used for the detection of HHV6 antigen in tissue sections.

Strong staining of the nuclei and cytoplasm of HHV6 infected, but not uninfected, J Jhan cells was seen with the anti-HHV6 monoclonal pool. A commercially acquired CMV MAb did not react with the HHV6 infected or uninfected J Jhan cells. This CMV- specific antibody was used, with equivalent sensitivity to in situ hybridization, for the detection of CMV infected cells in lung tissue from an AIDS patient. The anti-HHV6 monoclonal pool did not react with the CMV infected lung tissue. An anti-rubella haemagglutinin monoclonal, which was used as a negative control antibody, did not react with the HHV6 infected tissue culture cells or CMV infected tissue.

3c. 4. Detection of HHV6 DNA and protein in salivary gland tissues:

As part of an investigation into the role of saliva in the transmission of HHV6, salivary gland tissues were examined for evidence of HHV6 infection using DNA and protein detection techniques.

3c. 4. i. HHV6 DNA by nested PCR

Submandibular salivary gland tissues were obtained at postmortem from 5 individuals who had died from myocardial infarction. Nucleic acid was extracted from these tissue samples and tested for HHV6-specific DNA by nested PCR. High concentrations of salivary gland extract (containing approximately 5 pg of DNA) were inhibitory in the PCR reaction and when HHV6 plasmid was diluted in this DNA a 100 fold drop in sensitivity was noted. All salivary gland samples were positive for HHV6 DNA when diluted to contain approximately 0.05 - 0.5 pg DNA. The localization of HHV6 DNA and the detection of virus-specific proteins within salivary gland tissue sections was then attempted.

128 Figure 3c. 4

Photomicrographs of the fluorescence produced by 3 HHV6 monoclonal antibodies by indirect IF

A: Smooth staining with capping' of some HHV6 infected cells. B: Discrete, punctate staining C: Smooth staining throughout the nucleus and cytoplasm

129 3c. 4. ii. HHV6 DNA by in situ hybridization

Paraffin sections for in situ hybridization were obtained from salivary gland tissue taken as routine biopsies from 14 patients with benign tumours or stones (9 submandibular and 5 parotid glands). All tissues were screened for the presence of HHV6 and CMV DNA by in situ hybridization.

The HHV6 probe reacted strongly with submandibular and parotid salivary gland cells (Figure 3c. 5). All of the 9 submandibular glands contained strongly staining cells of both mucous and serous types. In mucous cells both nuclei and cytoplasm were stained whereas in serous cells staining was restricted to the nuclei. In 2 of the 5 parotid glands, acini of serous cells showed strong nuclear reactions with little or no cytoplasmic staining. Striated ducts in the 9 submandibular glands and the 2 parotid glands also showed strong staining for HHV6 DNA in the nuclei and cytoplasm of the columnar cells.

Seven of the 9 submandibular glands also showed nuclear and cytoplasmic staining of mucous cells for CMV-specific DNA. Serous cells did not react with the CMV probe and staining of striated ducts was either very weak or absent. The proportion of submandibular gland cells which stained for CMV DNA was much less than stained for HHV6 DNA. None of the 5 parotid glands contained cells which reacted with the CMV probe.

No staining of any of the tissues studied was seen with the control biotinylated plasmid.

3c. 4. iii. HHV6 protein by immunohistochemical staining

All salivary gland sections were also screened for HHV6 and CMV proteins by immunohistochemical staining. Eight of the 9 submandibular glands and the 2 parotid glands that were positive for HHV6 DNA showed scattered foci of cells staining for HHV6 protein ( Figure 3c. 6). The staining of mucous cells was limited to the cytoplasm and the staining of serous cells was limited to the nucleus. Striated ducts showed both nuclear and cytoplasmic staining for HHV6 protein.

Cytomegalovirus protein was detected in all of the 7 submandibular gland tissues which were positive for CMV DNA. Staining patterns were similar to that for the HHV6 antigen but no ductal staining was seen. No CMV protein was detected in the 5 parotid glands.

The control rubella MAb did not stain any of the salivary gland tissues.

130 HHV6 DNA-specific staining by in situ hybridization in submandibular and parotid salivary glands

A: Submandibular gland (X 2,000). HHV6 DNA present in serous acinus (S) and HHV6 diffusely throughout mucous cells in many acini (M). B: Parotid gland (X 2,000). HHV6 DNA localized within the nuclei (arrowed) of serous cells in many acini (S).

131 HHV6-specific protein staining with monoclonal antibodies in submandibular and parotid salivary glands

A: Submandibular gland (X 2,000). HHV6 antigens in the cytoplasm of mucous cells (M) and columnar cells of the striated duct (D). B: Parotid gland (X 2,000). HHV6 antigens in the nuclei of serous cells (S) and nuclei and cytoplasm of ductal cells (D).

132 3c. 5. Detection of HHV6 DNA by nested PCR in peripheral blood mononuclear cells and oropharyngeal samples

Blood and saline gargle samples were available from 12 healthy adults and were tested for HHV6 DNA by nested PCR. Initially a cellular fraction was obtained from the saline gargle by low speed centrifugation and DNA was extracted from the pelleted material and the supernatant. It soon became apparent that the majority of HHV6 in throatwashings (TW) was cell free and that ultracentrifugation of the sample prior to nucleic acid extraction gave a better indication of the number of individuals who were shedding HHV6 into the oropharynx. Peripheral blood mononuclear cells were obtained by Ficoll-Paque separation of whole blood and DNA was extracted from the pelleted cells. The concentration of DNA in the PBMC and TW samples was assessed by determination of the optical density at 260 nm and 0.1 - 2 pg of DNA was added to each PCR reaction.

All PBMC samples were HHV6 negative when < 0.5 pg of DNA was used in the PCR reaction but 5/12 (42 %) samples were HHV6 positive when 0.5 - 2 pg of PBMC DNA (the equivalent of 9 x 104 - 4 x 105 cells) was amplified. When HHV6 plasmid was diluted in 2 pg of HHV6 PBMC negative DNA there was no reduction in sensitivity and 5 copies of HHV6-specific sequence were detectable after two rounds of PCR.

When 0.1 - 0.5 pg DNA extracted from TW samples were added to the PCR reaction 6/12 (50 %) samples were HHV6 DNA positive after two rounds of PCR. Two TW samples were inhibitory when > 0.5 pg of DNA was added to the PCR reaction and when HHV6 plasmid was diluted in these samples a 10 - 100 fold reduction in sensitivity was noted.

The TW and PBMC samples were also tested for B-globin sequences using the PC03 and KM38 primers designed by Saiki et al. (1988). All samples were positive for B-globin DNA after first round PCR. When 0.1 - 2 pg of PBMC DNA was used in a single round PCR reaction all G-globin bands were of maximum strength when analyzed by agarose gel electrophoresis. When > 0.5 pg of DNA from the 2 inhibitory TW samples was amplified with G-globin primers the specific band was faint, all other TW samples gave a maximum strength band for 0.1 - 2 pg DNA. An example of a gel showing G-globin and HHV6 DNA amplification in tissue samples is given in Figure 3d. 3.

Multiple TW samples from 2 individuals taken over a period of 3 months were screened simultaneously for HHV6 DNA by nested PCR. Three of 6 samples taken from 1 individual and 5 of 6 samples from the other were HHV6 DNA positive after two rounds of PCR. It would seem from these results that either HHV6 is shed intermittently

133 into the oropharynx or that the concentration of virus in TW varies.

Nested PCR was compared with single round PCR with Southern blot confirmation for the detection of HHV6-specific DNA in TW and PBMC samples. All 24 samples from healthy individuals were screened for HHV6 DNA by Dr R. Gopal (National Institute for Medical Research, London) using the H6 and H7 primers and Southern blot confirmation, with concordant results to those obtained by nested PCR.

The ability to detect HHV6 DNA in TW or PBMC samples did not seem to be related to the presence of HHV6-specific antibody in serum. One individual was consistently negative for anti-HHV6 IgG by indirect IF and competitive RIA over a 3 year period but had detectable HHV6 DNA in TW samples taken during this time.

3c. 6. Discussion

Methods for the detection of HHV6 in tissue samples, peripheral blood and saliva were developed and used to identify potential sites of HHV6 replication and persistence. A high proportion of histologically normal submandibular and parotid salivary glands were found to contain HHV6-specific DNA and protein. This study was the first to identify a potential site of HHV6 replication relevant to transmission of the virus between normal healthy individuals. This observation was later confirmed by Krueger and colleagues (1990). As the HHV6 MAbs had not been characterized it was not possible to distinguish between HHV6 latent infection and active replication of salivary gland tissue. Krueger et al. (1990) used an HHV6 MAb which recognizes the p41 protein of HHV6 (Balachandran et al ., 1989). This protein has been partially characterized (Chang and Balachandran, 1991)

Two groups reported the frequent isolation of HHV6 from the saliva of healthy individuals (Pietroboni et al., 1988a, Harnett et al., 1990, Levy et al., 1990a) although others have failed to isolate the virus from throat swab or saliva samples (Kido et al., 1990, Lopez and Honess, 1990, Yoshiyama et al., 1990). When DNA extracts from TW samples (0.1 - 0.5 pg) were amplified by nested PCR 50 % of samples were found to be HHV6 DNA positive. The results for HHV6 amplification by nested PCR from the 12 TW samples was confirmed in another laboratory using single round PCR and Southern blotting. In two cases inhibition of the PCR reaction was seen when > 0.5 pg of TW DNA was added to the PCR reaction. Inhibitors of the PCR reaction which co-purify with genomic DNA have been reported previously (de Franchis et al., 1988). There have been three other reports of the detection of HHV6 DNA in oropharyngeal samples by PCR (Gopal et al., 1990, Jarrett et al., 1990, Kido et al., 1990).

134 It would seem from all these results that HHV6 is present in the oropharynx of a large proportion of healthy individuals and the source of this shed virus may be infected salivary gland tissue. The shedding of virus in the oropharynx seems to vary with time and is independent of serological status of the individual. Jarrett and colleagues (1990) also reported HHV6 'seronegative" individuals with detectable HHV6 DNA in their saliva.

The early isolates of HHV6 (Salahuddin et al., 1986, Tedder et al., 1987, Downing et al., 1987) were all obtained by culture of PBMC from patients with lymphoproliferative disorders or AIDS. Isolates of HHV6 have also been obtained by culture of PBMC from infants with roseola infantum (Yamanishi et al., 1988, Table 4. 2). Although many people have cultured PBMC from healthy adults only two groups have obtained HHV6 isolates from this source (Lopez et al., 1988, Daugherty et al., 1991, Table 4. 1).

It would seem that only a few circulating lymphocytes from healthy individuals carry HHV6 DNA and the detection of HHV6 in PBMC is dependent on the amount of DNA added to the PCR reaction. When nucleic acid extracted from approximately 1 X 105 PBMC was amplified by PCR, 42% of samples were HHV6 DNA positive. These results are comparable with another published report of HHV6 in peripheral blood samples from healthy adults (Jarrett et a l, 1990). Kondo and colleagues (1990, 1991) fractionated Ficoll-Paque separated PBMC from healthy adults into adherent (monocyte) and non-adherent (lymphocyte) cells. HHV6-specific DNA detected by PCR was associated mainly with adherent cells and the authors suggested that monocytes / macrophages may be a site of HHV6 latency.

135 3d. Detection of HHV6-specific DNA and protein in samples from patient groups

3d. 1. HHV6 DNA by in situ hybridization in postmortem tissues from cases of sudden infant death syndrome

Sudden infant death syndrome (SIDS or cot death) is defined as the death of an infant which is unexpected from clinical history and unexplained by postmortem investigation. The main histological changes in tissues from SIDS cases taken at autopsy are within the lung and consist of oedema, congestion and pleural hemorrhage. There may be many underlying causes of sudden death in an infant which makes study of a group of SIDS cases difficult. Much research interest has centered around the identification of infectious agents which may be responsible for this syndrome. Chlamydia (Schachter and Grossman, 1990), group B streptococcal (Baker and Edwards, 1990) and Shigella (Guerrant et al., 1990) infections have all been implicated in cases of sudden infant death. The hypothesis that SIDS may be a result of overwhelming viral infection has received some support. A number of groups have isolated enteroviruses from brain, lung and myocardium tissues from cot death cases (discussed in Cherry, 1990). Other viruses that have been implicated as aetiological agents in SIDS include CMV, respiratory syncytial virus and influenza virus A and B (Variend and Pearse, 1986).

Berenberg and colleagues (1949) first suggested that the causative agent of roseola infantum may be responsible for some cases of sudden infant death. The early acquisition of HHV6 and the association of this virus with roseola infantum (Yamanishi et al., 1988) led to speculation as to a potential role for HHV6 in SIDS.

Tissues from SIDS cases were autopsied at The London Hospital Medical School and were investigated for the presence of HHV6- and CMV-specific DNA by in situ hybridization. This was part of a larger study of viral infections associated with sudden infant death coordinated by Dr A. Coumbe (London Hospital Medical School, Coumbe et al., 1990). In each case the cause of death was unascertainable and fulfilled the definition of SIDS (average age 3.7 months, range 5 weeks to 7 months). Lung, thymus and spleen tissue were available from 30 and lymph node from 4 cases of sudden infant death. There was no HHV6 positive tissue available at the time of this study and so HHV6 infected and uninfected J Jhan cells, which were embedded in gelatin and wax, were used as controls for the in situ hybridization reaction. Postmortem tissue from an AIDS case with both CMV pneumonitis and Pneumocystis carinii infection was used as

136 a positive control tissue for CMV. All tissues were screened for specific DNA using biotinylated CMV, HHV6 and negative control probes as described in 3c. 2.

Control tissue gave a clear signal with little or no background staining. In the control lung tissue, positive staining was seen with the CMV probe in inclusion bodies of virus infected cells as shown in Figure 3d. 1, these cells were not stained with the HHV6 or negative control probes. The CMV-specific DNA probe and the negative control probe did not react with HHV6 infected J Jhan cells (Figure 3d. 2A). The infected J Jhan cells showed mostly cytoplasmic staining with the HHV6 probe but nuclear staining was seen in syncytia (Figure 3d. 2B). No positive staining for either CMV or HHV6 was seen in the SIDS material examined. As in situ hybridization was found to be more sensitive for the detection of HHV6 infected tissues than immunohistochemistry (see 3c. 4) it was not felt to be necessary to screen tissues for virus-specific proteins.

The results of this study did not indicate an association of HHV6 (and or CMV) infection with SIDS.

3d. 2. HHV6 DNA by nested PCR in tissues from patients with inflammatory bowel disease

Crohns disease (CD) and ulcerative colitis (UC) are chronic inflammatory bowel diseases (IBDs) of unknown pathogenesis. The highest incidence of both CD and UC is in young adults amongst the white populations of both North Western Europe and North America. The figures for Southern Europe, Eastern Europe and Japan are low (Langman, 1990) and CD and UC seem to be less common in black than in white populations.

Crohns disease may affect any part of the alimentary canal from the mouth to the anus although the terminal ileum and anal region are the sites most commonly associated with disease. All layers of the bowel wall are involved and a characteristic feature of CD is the presence of granulomas. Anal or perianal lesions are the most common early clinical manifestation of disease. When the terminal ileum is involved symptoms include abdominal distension and pain, diarrhoea and bleeding.

Ulcerative colitis is an inflammatory disease of the large bowel; the rectum and colon are most often involved. In uncomplicated UC, disease seems to be confined to the mucosa and only in a small number of cases are other layers of the bowel wall affected. Rarely, deep ulceration may extend into the muscle and serosa, resulting in perforation. Early clinical symptoms of UC are variable but most commonly include constipation associated with the passage of blood and followed by diarrhoea with blood, mucus, pus and excessive flatus.

137 CMV DNA-specific staining by in s itu hybridization in lung tissue from an AIDS patient

A classic owl's eye' inclusion in a type II pneumocyte is shown (X 4,(XX)). Staining for CMV DNA (red) is seen in both nucleus and cytoplasm, the tissue is counterstained with haematoxylin (purple). Figure 3d. 2 HHV6 infected tissue culture cells probed for virus-specific DNA by in situ hybridization

• * * #**

f ..... **■ *** i ; t f c - 4 -' *

■fir- W :J * - - V i r l 8ff

HHV6 infected J Jhan cells sectioned and screened for viral DNA using CMV (A) and HHV6 (B) specific probes (X 1,600). HHV6 staining is both nuclear and cytoplasmic, cells are counterstained with haematoxylin (purple). Both CD and UC are characterized by cycles of acute disease and subsequent remission. Although corticosteroids may reduce the severity of these symptoms treatment often becomes less effective which each relapse. Exacerbation of disease can be serious and even life-threatening (Langman, 1990). Perforated ulcers and eroded blood vessels may lead to severe haemorrhage and peritonitis. Many cases require surgical intervention, often many times.

Environmental factors such as smoking, drugs and diet have been associated with some cases of IBD (discussed in Heaton, 1990, Langman, 1990). Other research workers have suggested that CD and UC may be caused by an autoimmune mechanism. There is much evidence of immunological abnormalities in IBD which may be important in the perpetuation of the disease state (Jewell and Snook, 1990).

The search for infectious agents associated with CD and UC has focused on bacteria and viruses which may persist in the body and cause chronic inflammation (reviewed by Thayer, 1990). Antibody-mediated endothelial cell injury may account for the vasculitis which has been associated with CD and UC (Cines, 1989, Roberts et al ., 1989, Wakefield et ah, 1989). Classical CMV inclusions have been identified in mesenteric vascular endothelial cells of tissues from UC patients by in situ hybridization (Morton, 1985, Roberts et al ., 1989) and HSV has been shown to produce persistent infection of endothelial cells in vitro (Cines, 1989). Immunocompromised (particularly AIDS) patients may suffer from fulminant CMV colitis which has some similarities with UC and may be an exacerbation of this disease. One group (Farmer et al ., 1973) reported a higher prevalence and titre of CMV antibody in UC patients when compared with controls but the level of CMV antibody did not correlate with the duration and severity of illness. Other serological studies failed to show any association of active viral infections with IBD.

In view of the possible association of herpesvirus infections with IBD a study was undertaken to examine the prevalence of HHV6 in tissues from CD and UC patients. This was part of a larger study coordinated by Dr A. Wakefield (Royal Free Hospital School of Medicine) to investigate the pathogenesis of CD and UC. Resection and biopsy specimens of large intestine from 21 patients with UC (mean age 43 ± 17) and 29 patients with CD (mean age 39 ± 19) were available. Samples of large intestine were also obtained from patients who had had a resection due to malignant disease. Histologically normal intestine from these patients (mean age 61 ± 18) was used as a source of non-inflammatory control tissue in this study. Nucleic acid was extracted from all tissues by research workers in another laboratory. Tissue extracts were initially tested for HHV6-specific DNA by nested PCR. In order to ensure that the tissue derived DNA did not inhibit the PCR reaction 5 tissue samples from each group were also

139 tested for B-globin sequences as described by Saiki et al. (1988). When 0.1 pg of each DNA was used in the PCR reaction and the first round products were analyzed by agarose gel electrophoresis all B-globin bands were of maximum strength. An example of a PCR gel showing first round HHV6 and B-globin specific products for 3 intestinal DNA samples is given in Figure 3d. 3. Tissue samples which were negative for HHV6 were 'spiked" with the equivalent of 50 copies of HHV6 DNA. The HHV6 sequence was amplified from all these samples giving a weak first round and maximum strength second round band confirming that these nucleic acid extracts did not contain PCR inhibitors. First and second round HHV6-specific PCR products amplified from patients’ intestinal tissue extracts and analyzed by agarose gel electrophoresis are shown in Figure 3d. 4.

In order to interpret the nested PCR results for HHV6 in intestinal tissues and to determine any coinfection with the other human herpesviruses, nested PCR amplification of CMV, EBV, VZV and HSV 1 DNA was undertaken. Primers were either designed from available sequence data or adapted from published PCR methods. All PCR reactions were optimized in the same way as for HHV6 nested PCR (3c. 1.). Specific details of the primer design and optimization experiments for herpesvirus nested PCR are not relevant to this thesis but primer sequences, anneal temperatures and product sizes are given in Table 2. 1 (see also Wakefield et al ., 1992).

This study provided some interesting data on the prevalence of herpesviruses in large intestinal tissues. None of the extracts from 8 UC, 29 CD and 10 control tissues contained VZV or HSV-1 specific DNA and subsequent samples from IBD patients were not tested for these viruses. It was concluded from these results that VZV and HSV 1 are not involved in the pathogenesis of either UC or CD.

A summary of results for HHV6, CMV and EBV DNA by nested PCR in tissues from IBD patients is given in Table 3d. 1. There was no clear relationship between age, sex or corticosteroid treatment of patients and nested PCR status for HHV6, CMV or EBV DNA (Wakefield et al ., 1992).

Positive signals for HHV6 DNA were obtained from 16 / 21 (76 %) tissues from patients with UC, 13/29 (45 %) tissues from patients with CD and 18/21 (86 %) non inflammatory control tissues. Differences in prevalence of HHV6 DNA when these three groups were compared were not significant. This was the first identification of HHV6 infection of intestinal tissue. When tissue extracts were tested for CMV DNA by nested PCR the number of positive samples was significantly different when control and UC groups (6/21 compared with 17 / 21, X2 test p < 0.01) and when control and CD groups were compared (6 / 21 compared with 19 / 29, test, p < 0.05).

140 Figure 3d. 3

Single round PCR amplification of HHV6 and P-globin sequences in DNA extracted from intestinal tissue

1 2 3 4 5 6 7

H H - B

First round PCR products analyzed by agarose gel electrophoresis. Lane 1 = d>X174 / Hae III molecular weight markers. Products of PCR amplification from 1 control (lanes 2 and 3), 1 ulcerative colitis (lanes 4 and 5) and 1 Crohns disease (lanes 6 and 7) patient. Lanes 2. 4 and 6 = amplification products with HHV6 primers, lanes 3, 5 and 7 = amplification products with p-globin primers. H = HHV6 specific product (223 bp), B = p-globin specific product (167 bp).

I4l Figure 3d. 4

Nested PCR amplification of HHV6 DNA in tissue samples from inflammatory bowel disease patients

i B

First (A) and second (B) round PCR products analyzed by agarose gel electrophoresis. Lane 1 = X174 / Hae III molecular weight marker. All other lanes contain amplified products from ulcerative colitis or Crohns disease intestinal DNA. Arrow shows HHV6 specific PCR product (223bp 1st round, 173 bp 2nd round).

142 Table 3d. 1

Detection of HHV6 , CMV and EBV DNA by nested PCR in tissues from inflammatory bowel disease patients and controls

Number of tissues positive Patient for virus-specific DNA Group CMV HHV6 EBV

ulcerative 17/21 16/21 16/21 colitis (81%) (76%) (76%)

Crohn's 19/29 13/29 16/29 disease (66%) (45%) (55%)

non-inflammatory 6/21 18/21 4/21 control (29%) (86%) (19%)

There was a significant difference in the number of CMV positive samples when control and ulcerative colitis groups (6/21 compared with 17/21, X2test, p < 0.01) and control and Crohn's disease groups were compared (6/21 compared with 19/29, X2test, p < 0.05).

There was a significant difference in the number of EBV positive samples when control and ulcerative colitis groups (4/21 compared with 16/21, X2 test, p < 0.001) and control and Crohn's disease groups were compared (4/21 compared with 16/29, X2 test, p < 0.05).

143 Differences in the number of tissue extracts containing EBV DNA were also significant when control and UC groups (4/21 compared with 16 / 21, X2 test, p < 0.001) and control and CD groups were compared (4/21 compared with 16/29, X2 test, p < 0.05).

The high prevalence of EBV in inflammed tissues (CD and UC) when compared with controls may be explained by the accumulation of B-lymphocytes at the site of chronic inflammation. This argument does not hold for HHV6 since many of the non inflammatory control tissues were positive for HHV6 DNA by nested PCR.

In view of the unexpectedly high prevalence of HHV6 DNA in tissues from both UC and non-inflammatory control patients, data on the simultaneous presence of HHV6 with either or both CMV and EBV DNA in the same tissue extract were analyzed. Thirteen of 21 samples from UC patients (62 %) were positive for both HHV6 and CMV DNA compared with 8 /2 9 (28 %) samples from CD patients and 5/21 (24 %) control samples. The differences between UC and both CD and control samples were significant (X2 test, both p < 0.05) but the difference between CD and control samples was not significant. Thirteen of 21 samples from UC patients (62 %) were positive for both HHV6 and EBV compared with 7 / 29 (24 %) samples from CD patients and 2/21 (10 %) control samples. The differences between UC and both CD and control samples were significant (X2 test, both p < 0.01) but the difference between CD and control samples was not significant.

The association of two or more herpesviruses with UC was even more apparent when nested PCR results for full thickness resection tissue (not biopsy tissue) were analyzed. Ten of 12 (83 %) full thickness resection tissues from UC patients were positive for both HHV6 and CMV DNA compared with 5 /2 0 (20 %) resection tissues from CD patients and 2 / 9 (22%) resection tissues from controls. The differences between UC and CD and UC and control tissues were significant (X2 test, p < 0.001 and p < 0.05, respectively). Ten of 12 (83 %) full thickness resection specimens from UC specimens were positive for both HHV6 and EBV DNA compared with 8 / 29 (28 %) resection tissues from CD patients and 1 / 9 (11 %) resection tissues from controls. The differences between UC and both CD and control tissues were highly significant (X2 test, both p < 0.001)

The simultaneous presence of HHV6 and CMV and or EBV in tissues from UC patients (16 / 21, 76%) was much greater than in either tissues from CD patients (11 /29, 38%) or control tissues (6/21, 29%); these differences were significant (X2 test, p < 0.05 for each). The interaction of HHV6 with CMV and or EBV may be important in the pathogenesis of UC and requires further study.

144 3d. 3. HHV6 DNA by nested PCR in gut lymphomas

The lymphotropic nature of HHV6 and the high levels of HHV6 DNA in intestinal tissues led to speculation that the virus may be associated with the pathogenesis of lymphoid malignancies of the gut.

Extracted DNA from 24 low grade B-cell lymphomas associated with the lymphoid mucosa of the gut were provided by Prof. P. Isaacson (University College and Middlesex School of Medicine). These tissue extracts were tested for HHV6-specific DNA by nested PCR. Eleven of the 24 samples (46 %) contained detectable HHV6 DNA after two rounds of PCR.

In view of the extreme sensitivity of nested PCR and previous data showing that a large proportion of non-inflammatory intestinal tissues were HHV6 positive using this technique, these results were not thought to be significant. Four tissue samples from T- cell gastric lymphomas were all negative for HHV6 DNA by nested PCR.

3d. 4. HHV6 DNA by nested PCR in malignant tumours derived from nervous tissue

There had been reports of in vitro HHV6 replication in glial cells and glioblastoma cell lines (Tedder et al ., 1987, Ablashi et al, 1987). An investigation to determine any association of HHV6 infection with the pathogenesis of malignant tumours derived from nervous tissue was undertaken.

Extracted DNAs from 81 malignant tumours of the central and peripheral nervous system were provided by Prof. P. Isaacson and were tested for HHV6 DNA by nested PCR. Two of the 9 medulloblastomas and 1 of the 18 neuroblastomas tested were HHV6 positive by nested PCR. Other tumours (including glioblastomas and astrocytomas) were all negative for HHV6-specific DNA. Although the demonstration of HHV6 DNA in samples from malignant tumours of nervous tissue is interesting, as only 3 of 81 tissue extracts were PCR positive, the involvement of HHV6 in the pathogenesis of these malignancies is unlikely.

3d. 5. HHV6 protein by immunohistochemistry in tissues from Sjogren's syndrome patients

Sjogren's syndrome is an autoimmune disease affecting the salivary and lacrimal glands. The disease is characterized by benign or malignant lymphoid proliferation and destruction of affected glands. Proliferation of monocytic cells may also occur in other sites and patients suffering from SS have an increased risk of developing malignant lymphoma. The syndrome may occur independently or associated with other

145 autoimmune disorders such as rheumatoid arthritis. Fox and colleagues (1986, 1987) proposed that SS may result from an abnormal immune response to a ubiquitous virus.

The detection of HHV6 in salivary gland tissue led to an investigation of the prevalence of this virus in tissues from SS patients. Biopsy tissue from 10 patients with benign lymphoproliferative lesions of parotid salivary glands was provided by Prof. C. Scully (University of Bristol). Sections from these parotid gland biopsy and control tissues were stained for HHV6-specific proteins by immunohistochemistry. All of the parotid glands were positive for HHV6 protein but negative for CMV and rubella proteins by this method. The staining pattern of serous and columnar cells was similar to that described previously (3c. 4. iii. and Figure 3c. 6B). The mononuclear cell infiltrate in each case also showed HHV6-specific protein staining (Figure 3d. 5A), but these cells did not stain with the CMV-specific MAb (Figure 3d. 5B).

As all tissues were positive for HFIV6-specific protein it did not seem to be necessary to demonstrate the presence of virus by the more sensitive in situ hybridization technique. Krueger and colleagues (1990) tested lip biopsy specimens from SS patients and healthy controls for evidence of HHV6 infection. All tissues were positive for HHV6 antigen using the P41 MAb described by Balachandran and coworkers (1989). These results suggest that HHV6 may be associated with the pathogenesis of SS (and perhaps other autoimmune diseases).

3d. 6. Discussion

Methods for the detection of HHV6-specific DNA and protein were applied to the identification of virus infected tissues from patients.

No evidence was found for the association of HHV6 (or CMV) infection with sudden infant death. Although in situ hybridization seemed to work well for the detection of viral DNA in control cells and tissues no specific staining was seen in lung, thymus, lymph node or spleen from cases of SIDS. If an overwhelming infection with either virus was the cause of a significant proportion of SIDS cases it would seem likely that viral DNA would have been detected in at least some of the tissues screened.

Nested PCR was found to be a very useful technique for the rapid (and sensitive) screening of extracted tissues for the presence of viral DNA. A high proportion of large bowel tissue extracts from IBD and from non-inflammatory control patients was found to contain HHV6-specific DNA by this method. The identification of HHV6 nucleic acid sequences in these tissues was unexpected and had not previously been reported.

In order to try and interpret these results, a similar technology was used for the

146 Parotid gland lymphoproliferative lesions screened for

HHV 6 - and CMV-specific proteins using monoclonal antibodies

A r • J jA A g j I Z ¥ 4 s *

V S ' 1 • • » I * •> e V « v \ V * V . « ‘ y

. * f) M <1 *'•

• * « • % * * •»% - fcl' "»

.I. * *4 v ♦> /• r

A: HHV6 specific staining of mononuclear cells (brown) is mainly cytoplasmic. B: There is no staining of mononuclear cells with CMV monoclonal antibody. Tissues were counterstained with haematoxylin (purple) X 1,600.

147 detection of CMV, EBV, VZV and HSV-1 DNAs in these tissues. Varicella-zoster virus- and HSV 1-specific DNA were not detected by nested PCR in the tissue extracts from patients or controls. Nucleic acid sequences specific for CMV and EBV were found in the majority of tissue samples from patients with CD and UC but only in a minority of control tissues. The presence of these viruses in patients' samples may be a secondary effect due to infiltration of infected cells as part of the inflammatory response. This would not, however preclude an association of these viral infections with disease. The presence of HHV6 in large bowel tissue was not related to the inflammatory response associated with bowel disease since specific DNA was detected in a high proportion of control tissues. The detection of HHV6-, CMV- and EBV- specific DNA in tissue extracts did not seem to be related to the age, sex or corticosteroid treatment of patients.

The prevalence of coexistant HHV6 with either or both CMV and EBV was significantly higher in UC patients than in CD and control patients. Interactions between these viruses may be important in the pathogenesis of UC. In order to prove an aetiological role for these viruses in UC, the in situ localization of virus-specific DNA and protein with diseased areas of tissue would have to be demonstrated. Tissue sections were not available from enough patients and controls to make this part of the study worthwhile at this time.

Two research groups used Southern blotting to investigate the possible role of HHV6 in the pathogenesis of lymphoid malignancies (Josephs et al., 1988a, Jarrett et al., 1988). Over 150 tissues were screened by each research group and 5 patients with HHV6 positive lymphomas were identified. These results suggested that HHV6 may be associated with some B-cell lymphomas. As no similar studies on tissues from individuals without lymphoid malignancies and no in situ localization of specific DNA was reported, these results were difficult to interpret.

When B-cell gut lymphoma tissues were screened for the presence of HHV6 DNA by nested PCR almost half contained virus-specific DNA. In view of the extreme sensitivity of nested PCR and the high prevalence of HHV6 in normal intestinal tissues these results would suggest that HHV6 is only rarely, if ever, associated with the pathogenesis of gut lymphomas. One research group (Jarrett et al., 1988) reported the detection of HHV6 DNA in a gastric biopsy from a patient who had had a B-cell gastric lymphoma. In view of the identification of HHV6 in normal intestine it would be difficult to prove that this virus had more than an incidental role in the pathogenesis of this lymphoma.

148 The detection of HHV6 DNA by nested PCR in 1 of the neuroblastoma and 2 of the medullablastoma tissues screened could indicate an aetiological role for HHV6 in the pathogenesis of these malignancies. Many more samples would have to be tested for the association of HHV6 with tumours of nervous tissue to be confirmed. Again the localization of HHV6 within these tissues would help to distinguish between the incidental detection of a ubiquitous virus and the identification of a viral infection which may be important in the disease process.

Buchbinder and coworkers (1988) reported the use of PCR to screen tumours for the presence of HHV6 DNA. Twenty of 23 tumours were HHV6 positive by this method, only 1 was positive by Southern blotting (reported by Josephs et al., 1988). These results demonstrate the extreme sensitivity of PCR for the detection of viral sequences and indicate that such positive results should be interpreted with caution. Torelli et al. (1991) reported that 3 of 25 Hodgkin's lymphoma tissues were HHV6 DNA positive by PCR.

The common finding of HHV6 DNA and protein in histologically normal salivary gland tissues makes the association of HHV6 with SS difficult to prove. Two of the patients with HHV6 DNA positive lymphomas previously reported (Josephs et al., 1988a, Jarrett et al., 1988) suffered from SS. In this study all salivary glands from SS patients were positive for HHV6 proteins by immunohistochemistry. Serous, columnar and mononuclear cells were stained for HHV6 proteins using specific MAbs. These cells were not stained for CMV or rubella antigens. The detection of specific proteins in infiltrating lymphoid cells may indicate an association of HHV6 with the pathogenesis of SS.

149 3e. A longitudinal study of HHV6 in pregnancy

3e. 1. Introduction and aims of this study

Clinical symptoms associated with HSV infection are frequently noted during pregnancy (Whitley, 1990) but intrauterine infection of the foetus with HSV is thought to be rare (Whitley, 1990a). Neonatal HSV 2 infection may occur at birth as a result of cervical shedding of the virus and occasionally causes a generalized infection (Stagno and Whitley, 1985a, Whitley, 1990a). Primary VZV infection during pregnancy may lead to intrauterine infection and infected neonates may suffer from disseminated VZV (Stagno and Whitley 1985a, Heath and Kangro, 1990). Effects on the foetus when the mother seroconverts for EBV during pregnancy are thought to be rare (Fleisher and Bolognese, 1983, Stagno and Whitley, 1985, Crawford and Edwards, 1990). Congenital CMV infection may occur as a result of primary infection, reactivation or reinfection in the mother. Some studies have correlated impaired cell mediated immunity (CMI) to CMV in the mother within utero infection of her offspring (Rola-Pleszczynski et al ., 1977, Starr et al., 1979, Reynolds et al., 1979, Gehrz et al., 1981, Stem et al., 1986). Reactivation of CMV in pregnant women may result in renewed shedding of infectious virus (Stagno et al., 1975, Pass et al., 1982, Pass, 1985) and some infants acquire their CMV at birth due to cervical shedding of the virus. Horizontal transmission of CMV from mother to child may occur in the perinatal period by infected breast milk or saliva (Stagno et al., 1980, Dworsky et al., 1983, Alford and Britt, 1990).

Initial serological studies established that a large proportion of children are infected with HHV6 during the first year of life. This together with the fact that roseola infantum occurs as sporadic cases and not in epidemic form (Stammers, 1988) suggested that the virus is transmitted from previously infected adults to young children. There was already some serological evidence based on IgM antibody detection for increased replication of HHV6 during pregnancy (see 3a. 8.) and so it seemed that young children might have acquired HHV6 from their mother. None of a group of cord blood samples was anti-HHV6 IgM positive arguing against interuterine infection (3a. 7., Farr et al., 1990). One research group attempted culture of HHV6 from 50 endocervical swabs, with negative results (Harnett, et al., 1990). Takahashi and colleagues (1988) found no evidence for transmission of HHV6 by breast milk and in view of the evidence for replication of HHV6 in salivary glands transmission by saliva seemed likely.

A longitudinal study of HHV6 in pregnancy was undertaken in collaboration with obstetricians and virologists at The Queens University, Belfast. The aims of this study

150 were to measure parameters of HHV6 infection in samples from pregnant women and to investigate the potential role of saliva in the transmission of HHV6 to neonates.

3e. 2. Sample collection and analysis of data

Seventy women (mean age 27 ± 6 years) were enrolled into the study and serum, PBMC and TW samples were taken at various times during pregnancy and peuperium. The first samples were taken during the first or second trimester and then multiple samples were taken late in pregnancy and during the postnatal period. All but the first sample from each individual were coded so that it was not possible to match up samples from the same woman until all the results were available. A total of 301 sera, 301 PBMC samples and 298 TW samples were taken and, on average, samples from 4 different time points were available for each pregnant woman. Sera were tested for anti-HHV6 IgM by indirect IF and for anti-HHV6 IgG by indirect IF and competitive RIA. Throatwashings and PBMC samples were screened for FIHV6-specific DNA by nested PCR.

Results for samples from all 70 women in the study group were analyzed together in order to identify any changes in HHV6 antibody and HHV6 DNA prevalence or level. Antenatal samples were arranged into 6 groups (< 13, 13 - 18, 19 -24, 25 - 30, 31 - 36 and 37 - 42 weeks gestation) and postnatal samples into 2 groups (1- 6 and > 6 weeks postpartum) for this analysis. Samples were taken from 8 women during the first trimester (<13 weeks gestation group, average 11 ± 1 week) and from 15 women at > 6 weeks postpartum (average 20 ± 6 weeks). Where multiple samples from an individual fell within the same group only the sample taken closest to the middle time point was included.

Fifty seven women had samples taken from 4 or more time points during pregnancy and the early postnatal period. Results from these women were used to assess the changes in HHV6 antibody and DNA within an individual.

3e. 3. HHV6 antibody in sera from a group of 70 pregnant women

Results for HHV6 antibody in sera from the 70 pregnant women are given in Table 3e. 1 and Figure 3e. 1.

There was a drop in the prevalence of anti-HHV6 IgG by IF when sera taken during the first trimester (<13 weeks gestation) were compared with those taken in the second trimester (13 - 18 and 19 - 24 week groups). Analysis of the group as a whole did not identify a clear relationship between anti-HHV6 IgG by indirect IF and weeks gestation or postpartum. The presence and level of anti-HHV6 by competitive RIA also varied

151 oo c m 2/9 a; 2/8

3/13 CN 5/15 total 10/20 14/40 24/49 (N os positive £ +-4 a C3 0/8 a by indirect IF i-H 1/15 5/20 5/58 2/13 7/40

OX) Anti-HHV6 IgM 11/49 strong

a> positive t- a o i> 6/8 7/9

O 7/13 total 12/15 13/20 29/49 40/58 22/40

P. positive 3 O L* OX)

eo Anti-HHV6 3/9 5/8 8/20 2/13 4/15 10/49 10/40 21/58 strong by by competitive RIA o positive * so Lm a> 5/8 4/9 5/13 total 11/20 11/15 19/40 37/58 26/49 positive >* T3 O 0 0 CO ON pQ ^H

by by indirect IF r-H r-H 5/20 4/15 9/40 i—H Anti-HHV6 IgG 17/58 12/49

G strong

co positive a> so ro O) X Id NOi § c

3 X >6

<13 1—1 to 19-24 13-18 31-36 37-42 CO 25-30 o weeks Q postnatal H gestation

152 Figure 3e. 1

HHV6 antibody in sera from a group of 70 pregnant women 100 1 80 -

<13 13-18 19-24 25-30 31-36 37-42 1-6 7-25 weeks antenatal weeks postnatal 100 n

<13 13-18 19-24 25-30 31-36 3 7^ 2 1-6 7-25 weeks antenatal weeks postnatal 100 -| O

<13 13-18 19-24 25-30 31-36 3 7 ^ 2 1-6 7-25 weeks antenatal weeks postnatal

total positive ■ strong positive

153 throughout pregancy and the early postnatal period. There was a drop both in the prevalence of anti-HHV6 by competitive RIA and in the number of sera giving > 60 % inhibition of label bound when first and second trimester sera were compared. The G-test indicated that there was no clear association between the prevalence and level of serum anti-HHV6 by indirect IF or competitive RIA and weeks gestation / postpartum (Table 3e. 2).

A large number of sera in the pregnant women study group contained anti-HHV6 IgM by indirect IF. Fifteen HHV6 IgM antibody positive sera from pregnant women were fractionated by sucrose density centrifugation prior to antibody screening. In all cases the anti-HHV6 IgM fluorescence was associated with the IgM fraction of the serum. The basal level of anti-HHV6 IgM in first trimester samples was difficult to assess from this study since the first group (<13 weeks gestation) was so small. Two of the 8 samples in this group (25 %) gave weak positive reactions for anti-HHV6 IgM by indirect IF. These results are in marked contrast to those found previously when 124 12 - 14 week gestation sera were assayed for HHV6 antibody, 4 of these samples (3 %) contained HHV6 IgM antibody (see 3a. 8). The presence and level of serum anti-HHV6 IgM varied throughout pregnancy with the highest prevalence in sera taken during the third trimester. Although there was no clear association between HHV6 IgM antibody and weeks gestation or postpartum in this study (G-test, Table 3e. 2) the detection of virus-specific IgM antibody was in contrast to results obtained previously for sera from 43 female blood donors, none of which contained anti-HHV6 IgM (see 3a. 8). These results indicated that there are changes in the serological markers of HHV6 infection during pregnancy and postpartum.

3e. 4. HHV6 antibody in sera from 57 women followed through pregnancy

By IF, 20 of the 57 women (35 %) for whom multiple sera were available were anti- HHV6 IgG positive and 11 (19 %) were anti-HHV6 IgG negative throughout the course of this study. By RIA, 28 women (49 %) were positive and 9 women (16 %) were negative for anti-HHV6 during this time period. Only 2 women were anti-HHV6 IgG negative by both IF and RIA during pregnancy and the early postnatal period. Fifteen women (26 %) seroconverted for anti-HHV6 IgG by IF, 13 women (23 %) seroconverted for HHV6 antibody by RIA and 3 women (5 %) seroconverted for anti- HHV6 by both EF and RIA. Sera from the remaining 16 women fluctuated between anti- HHV6 positive and negative by IF and or RIA over the study period.

Seven of the 57 women (12 %) were HHV6 IgM positive and 17 (30 %) were anti- HHV6 IgM negative throughout the course of this study. Twenty-two women (39 %)

154 Table 3e. 2

Statistical analyses of HHV6 antibody prevalence in sera from women taken at various times during pregnancy and the postnatal period

p ** HHV6 antibody G *

Anti-HHV6 IgG by indirect IF: total positives 6.74 > 0 .1 0 strong positives 4.75 > 0 .5 0

Anti-HHV6 IgG by competitive RIA: total positives 5.89 > 0 .5 0 strong positives 9.58 > 0 .0 5

Anti-HHV6 IgM by indirect IF: total positives 6.61 > 0 .5 0 strong positives 8.65 > 0 .1 0

* G value calculated using data from Table 3e. 1 in an R X C (8 X 2) test of independence (G-test).

** P values were read from a X2 distribution table (7 degrees of freedom).

The presence and level of anti-HHV6 in serum samples was independent of weeks gestation / postpartum.

155 seroconverted for anti-HHV6 IgM during pregnancy or the early postnatal period, 16 of these women (73 %) also seroconverted for anti-HHV6 IgG by IF and or RIA. Eleven women (19 %) fluctuated between anti-HHV6 IgM positive and negative during pregnancy.

These results confirmed that some women had changes in anti-HHV6 reactivity during pregnancy. In view of the high prevalence of HHV6 IgG antibody in adult groups it would seem more likely that these changes were due to HHV6 reactivation and not primary infection.

3e. 5. HHV6 DNA levels in peripheral blood mononuclear cells and throatwashings from a group of 70 pregnant women

Peripheral blood mononuclear cells and TW samples from pregnant women were screened for HHV6 DNA by nested PCR amplification, results are given in Table 3e. 3 and Figure 3e. 2.

From results obtained previously for non-pregnant adults it was clear that 2 pg of PBMC extracted DNA could be added to a PCR reaction without inhibitory effects. In order to confirm that this was also the case for samples from pregnant women 2 pg of DNA from each of 20 PBMC extracts taken at various times during pregnancy and the postnatal period were amplified using p-globin-specific primers. After a single round PCR reaction all p-globin bands were of maximum strength. When sufficient DNA was available, PBMC samples that did not contain detectable HHV6 DNA were spiked with the equivalent of 50 copies of HHV6 sequence and then reamplified. None of these PBMC samples contained inhibitors of the PCR reaction. Peripheral blood extracts containing HHV6-specific sequences were diluted to contain 0.5 pg of DNA and were reamplified. In no case did the strength of a first round HHV6 band increase nor did a first round negative sample become positive by dilution of the DNA.

When TW samples from non-pregnant adults were screened for HHV6-specific DNA by nested PCR two samples were inhibitory when > 0.5 pg of DNA was added to the reaction mixture. In order to ensure that TW samples from pregnant women were not inhibitory to the PCR reaction 0.5 pg of each of 20 TW extracts taken from women at various times during pregnancy and the postnatal period were screened for p-globin DNA as described for PBMC samples. All gave a maximum strength band of the expected size after a single round of PCR. When sufficient DNA was available TW samples were diluted and 0.1, 0.2, 0.3 and 0.5 pg of each DNA extract was amplified using HHV6-specific primers. In no case did the strength of a first round HHV6 band increase nor did a first round negative sample become positive by dilution of the DNA. Throatwashings which did not contain detectable HHV6 DNA by nested PCR were

156 Table 3e. 3

HHV6 DNA by nested PCR in peripheral blood mononuclear cells and throatwashings from a group of 70 pregnant women

PBMC TW weeks gestation First round Second round First round Second round positive positive positive positive

<13 0/8 3/8 0/8 4/8

13-18 2/40 13/40 5/39 24/39

19-24 3/13 7/13 4/13 8/13

25-30 2/9 6/9 5/9 8/9

31-36 0/20 1/20 6/20 18/20

37-42 0/49 10/49 14/49 42/49

1-6 0/58 11/58 15/58 51/58 postnatal >6 0/15 3/15 2/15 12/15 postnatal

157 mononuclear cells and throatwashings from a from throatwashings and cells mononuclear HHV Figure 3e. 2 3e. Figure

% % TW positive % PBMC positive 0 1 100 100 6 0 - 80 DNA by nested PCR in periperal blood periperal in PCR nested by DNA - group of 70 pregnant women pregnant 70 of group <13 <13

13-18 13-18 ■ First round positive round First ■ positive round Second E3 weeks antenatal weeks postnatal weeks antenatal weeks weeks antenatal weeks postnatal weeks antenatal weeks

19-24 19-24

25-30 25-30 158

31-36 31-36

2 ^ 7 3 ^2 7^ 3

1-6 1-6

7-25 7-25

spiked with 50 copies of HHV6 sequence. No inhibition of the PCR reaction was noted when 0.5 pg or less of TW DNA was added to the PCR reaction.

These preliminary experiments indicated that any change in HHV6 DNA prevalence or level in PBMC or TW samples during pregnancy was unlikely to be due to the presence of inhibitors of the PCR reaction.

None of the PBMC samples taken from pregnant women during the first trimester contained detectable HHV6 DNA after a single round of PCR. Three of the PBMC samples taken at < 13 weeks gestation (38 %) were HHV6 DNA positive after two rounds of PCR. These results are similar to those found previously for non-pregnant adults (see 3c. 5.). There was an increase in the prevalence of HHV6 DNA in PBMC samples taken at 13 - 30 weeks gestation. Many of the PBMC samples taken at 19 - 30 weeks gestation contained detectable HHV6 DNA after a single round of PCR. The increase in prevalence and level of HHV6 DNA in peripheral blood may be related to the drop in HHV6 antibody noted for sera taken during the second trimester. Late in pregnancy (> 30 weeks gestation) there was a drop in HHV6 DNA prevalence in peripheral blood samples. An association between prevalence of HHV6 DNA in PBMC and weeks gestation was confirmed using the G-test (Table 3e. 4).

None of the TW samples taken from pregnant women during the first trimester contained detectable HHV6 DNA after a single round of PCR. Four of the TW samples taken at < 13 weeks gestation (50 %) were HHV6 DNA positive after two rounds of PCR. These results are very similar to those found previously for non-pregnant adults (see 3c. 5.). There was an increase in the prevalence of HHV6 DNA in TW samples with advancing gestation. Many TW samples taken during the second and third trimesters contained detectable HHV6 DNA after a single round of PCR. Fifty-one of 58 (88 %) TW samples taken during the early postnatal period contained detectable HHV6 DNA by nested PCR. There was an association between % prevalence of HHV6 DNA in TW samples and weeks gestation (Table 3e. 4). It would seem from study of the relevant histogram (Figure 3e. 2) that the trend is for increased shedding of virus into the throat as pregnancy progresses.

3e. 6. HHV6 DNA in peripheral blood mononuclear cells and throat washings from 57 women followed through pregnancy

Seventeen of the 57 women (30 %) followed through pregnancy had no detectable HHV6 DNA in PBMC samples taken at anytime during the course of this study. One or more of the PBMC samples from the rest of the women contained HHV6 DNA. Twelve women (21 %) had detectable HHV6 DNA in PBMC samples taken early in pregnancy but those taken later and during the postnatal period were negative for HHV6 DNA.

159 Table 3e. 4

Statistical analyses of HHV6 DNA prevalence in peripheral blood mononuclear cells and throatwashings taken at various times during pregnancy and the postnatal period

p ** HHV6 DNA by nested PCR G *

TW samples: first round positives 13.31 > 0 .0 5 second round positives 15.28 < 0.01

PBMC samples: first round positives 16.44 < 0 .0 5 second round positives 20.29 < 0.01

* G value calculated using data from Table 3e. 3 in an R X C (8 X 2) test of independence (G-test).

** P values were read from a X2 distribution table (7 degrees of freedom).

The prevalence of HHV6 DNA in TW and PBMC samples was not independent of weeks gestation / postpartum.

160 Twenty-eight women (49 %) had detectable HHV6 DNA in PBMC samples intermittently during the study period.

All of the 57 women had detectable HHV6 DNA in one or more TW samples taken during the course of this study. Of these, 22 women (39 %) had detectable HHV6 DNA throughout. Twenty-six women (46 %) had a significant increase in the amount of HHV6 shed into the throat as pregnancy progressed. In most cases the high level of HHV6 DNA shed into the throat persisted into the postnatal period. In some women, where > 20 week postnatal samples were available, the level of HHV6 DNA peaked late in pregnancy or during the postnatal period and then dropped. Nine women had detectable HHV6 DNA in TW intermittently during the course of this study.

3e. 7. A summary of results for HHV6 in pregnancy

1) The prevalence and titre of HHV6 specific IgG dropped in the second trimester and then increased back to the levels found in female blood donors by the third trimester and postpartum. A small proportion of women had a rise in anti-HHV6 IgG level as pregnancy progressed.

2) At some time during the antenatal period sera from the majority of pregnant women contained HHV6 IgM antibody. The greatest proportion of anti-HHV6 IgM positives was found in samples taken late in pregnancy.

3) HHV6 DNA was detected at some point in peripheral blood samples from most women during pregnancy or the postnatal period. The proportion of women with detectable HHV6 DNA in PBMC dropped significantly during the third trimester and remained at low level during the postnatal period.

4) The proportion of women who had HHV6 in the throat and the amount of HHV6 shed into TW increased as pregnancy progressed. The high levels of virus DNA in TW were maintained into the postnatal period.

3e. 8. Discussion

On balance, the evidence is for increased HHV6 shedding into the oropharynx during pregnancy together with serological evidence of HHV6 reactivation. Multiple samples were available from most women which allowed changes in HHV6 antibody and DNA to be assessed within an individual. Results for the whole study group were shown to reflect the changes in HHV6 within an individual.

The anti-HHV6 IgG results for the study group were fairly constant throughout pregnancy except that there was a drop in anti-HHV6 IgG prevalence and titre in the

161 second trimester. This may not be an HHV6 specific phenomenon since a drop in total serum IgG with advancing gestation has been reported (Maroulis et al., 1971).

Three of the 57 women for whom 4 or more sera were available seroconverted for anti- HHV6 IgG by both IF and RIA during the course of this study. Two of these women also seroconverted for anti-HHV6 IgM, the other had no detectable HHV6 specific IgM in any of the sera tested. These are thought to be examples of reactivation of endogenous HHV6 and not primary infection since all three women had detectable HHV6 DNA in TW samples taken early in pregnancy. It would seem that although these women were HHV6 infected their level of serum anti-HHV6 was very low. Many of the women followed through pregnancy and postpartum were anti-HHV6 IgM positive at some time during the antenatal period, these may also be examples of HHV6 reactivation.

The majority of women had only intermittently detectable HHV6 DNA in PBMC samples taken during the course of this study. This was most apparent when multiple samples from the same woman were taken within a short space of time. Late in pregnancy and during the postnatal period the prevalence of HHV6 DNA in PBMC samples was lower than that in the first trimester.

The majority of pregnant women in this study shed HHV6 into the throat late in pregnancy and during the postnatal period. Many of these women had detectable levels of HHV6 DNA in TW samples after one round of PCR. The increase in shedding of HHV6 as pregnancy progresses may be due to localized viral replication. Previous work (see 3c. 4.) has shown that submandibular and parotid salivary glands may be a site of HHV6 persistence and the source of virus in oropharyngeal secretions. Reactivation and increased shedding of HHV6 may be as a result of hormonal stimuli, a phenomenon which has been reported for CMV and EBV (discussed in Pass, 1985, Stagno and Whitley, 1985).

Studies of CMV isolation rates in pregnant women have been carried out by other research laboratories. Two groups reported an increase in cervical and urinary shedding of CMV with advancing gestation (Reynolds et al., 1973, Stagno et al., 1975). In both studies CMV excretion usually began in late gestation and continued into the postnatal period and there was no change in CMV antibody status to coincide with viral shedding. Stagno and colleagues (1975) reported no difference in CMV cervical shedding when comparing pregnant and non-pregnant women and suggested that the apparent increase in secretion of CMV during the third trimester was due to suppression of viral shedding early in pregnancy. This was not thought to be the case for secretion of HHV6 since significantly more pregnant than non-pregnant women shed HHV6 in TW. Reynolds

162 and colleagues (1973) extended their isolation studies to PBMC and throat swab samples from the women who excreted CMV in urine or cervical fluid. There was no evidence for CMV viraemia in these women and only 2 of 56 women excreting CMV in urine or cervical fluid also shed virus into the throat.

Most infants who acquire CMV during the neonatal period do so at birth by infected cervical secretions or in the early postpartum period by infected breast milk (Stagno et al ., 1980, Dworsky et al ., 1983, Stagno, 1990). When a group of unselected postpartum women were examined for evidence of CMV excretion by viral isolation 32 % were found to shed vims in breast milk, 10 % in vaginal secretions, 7 % in urine and 2 % in saliva (Dworsky et al., 1983). Takahashi and colleagues (1988) reported that transmission of HHV6 by breast milk was not common since babies who were bottle fed suffered from roseola infantum with the same frequency as those who were breast fed. It would seem that many more women shed HHV6 in saliva than shed CMV in cervical secretions, breast milk, urine or saliva during late pregnancy and the early postnatal period. This may account for the greater incidence of HHV6 compared with CMV infections during the first year of life. Perinatal CMV infection is often asymptomatic and may occur in the presence of maternal antibody (Reynolds et al ., 1973, Kumar et al., 1984). Many more infants acquire HHV6 than are accounted for by the incidence of roseola infantum and so it would seem that most HHV6 infections are mild and may be clinically inapparent. As most children acquire HHV6 during the first year of life some HHV6 infections must also occur in the presence of appreciable levels of maternal antibody.

163 Section 4: General Discussion

The main aims of this thesis were to determine the prevalence of HHV6 infection in healthy individuals and to identify any diseases associated with primary infection and reactivation of this virus. Many other research groups were also working on this newly isolated herpesvirus during the same time period and the results given in this thesis need to be considered in the context of other published reports.

4a. Isolation of HHV6

The original isolates of HHV6 were obtained from the peripheral blood of patients with lymphoproliferative disorders, many of whom were also infected with HIV (Salahuddin et al., 1986, Tedder et al., 1987, Downing et al., 1987). Since that time there have been many more isolations of HHV6 from HIV infected and uninfected adults (Table 4. 1). Although many people have tried to isolate the virus from PBMC of healthy individuals there are only two reports of such isolations in the literature (Lopez et al ., 1988, Daugherty et al., 1991) and these viruses have not been extensively characterized. It would seem that HHV6 is not normally present in the circulating lymphocytes of healthy adults in a form that can easily be cultured.

In 1988, a collaborative study group reported the isolation of HHV6 from PBMC of patients with CFS (Ablashiet al., 1988, Straus, 1988). Further isolates of HHV6 have since been obtained from PBMC of patients with symptoms of prolonged fatigue (Luka et al., 1990, Daugherty et al., 1991, Josephs et al., 1991).

At around the time that serological studies were undertaken to determine the prevalence of HHV6 in the adult population, Pietroboni and coworkers (1988a) reported the isolation of HHV6 from the saliva of 9 healthy individuals. This group later repeated this work and confirmed their results (Harnett et al., 1990). These were the first reports of isolation of virus indicating a possible mode of transmission of HHV6 between healthy individuals. Since this time there have been further reports of HHV6 isolation from the saliva of healthy and HIV infected individuals (Levy et al., 1990, 1990a) although other groups have failed to isolate HHV6 from throat swab or saliva samples of healthy individuals (Kido et al, 1990, Yoshiyama et al., 1990, Lopez and Honess, 1990).

The biological properties of HSV, VZV, CMV and EBV led to speculation that HHV6 may cause disease in immunosuppressed individuals. Isolates of HHV6 have been obtained from PBMC samples from recipients of liver, kidney or bone marrow

164 Table 4.1 HHV6 isolation from adults 55 'o O C/3 ■4—* G o cd cd TO cd C/3 G O G a •i-H ■s3 G D -G NO C/3 a.s <-> - 2 & X c/3 CX cd (N ^ O no 8 o vo CJ 4 0 0 0 O o o o v C/D PQ S -G o n o n o n o n Q > h 4-* cd cd G

U < ,(D O nO (N 4 ^ • c CX PQ S o o o n < a HrH • 53 > H - r 53 HH > -1 4 4—» 4-^ 0 c •I—i g cd 0 O G G W) CX cd N O CI O & a U • rH <4-4 O o o O O > nO <4-1 CX PQ S 53 HH 53 CH > TJ © 4—> -* 4 4—» cd CD O (X 165

o o o o 3 • N o O o o o < J * ■Jr* a ^ ^ cd <3 cd c« C/D • ^4 T3 53 l-H > H 2 cd G o CX cd CD G i PQ cd G U <4-4 CX PQ S O o o o o . nO •rH £ < -O 53 H-4 > CX 4—> 4^» 4-^ h C cd cd OD cd w G - m TO CX _ Jo -4 4 cd 3 G >4 ‘4 ^ 04, , 4 0 , 4 0 , 4 0 CJ CJ CJ i s PQ PQPQ s s 3 1 ,CD -13 « T3 .S'! 'G 03 ’—11 04 "T-Sd ■S c > 3 W S 53 a® > —i t—( (-4 CD ^ • H cd CD a cd s —■s 3 O OO O PQ a 4-4 CD CD 1-4 <4D PQ n n C

u O O O O 4 ^ • <4—1 PQ s < o n 4 r • H < CX 3 5 HH > -4 4 4—» W o 00 cu e ID cx cd G CD G n n CJ »-( Table 4.1 continued HHV6 isolation from adults £ O 3 / 1 o cd o Q • ^H 4 £ u 4-» *«H O -> + C o cd

V O 4> h

o 3 3 c <& .a§ 3 12'-3 T3 o t=J ^ a fa g* 0 o 00 fa 2 2 fa E « rt * » > «* o rj C - c c - n 2 2 cd 64

, N O N O O u (N PQ O S •rH •rH fa 4-1 d c 2 6 tn ^ , W a <41CD O C o i v n

h c > U fa 3 PQ D i ( a f V3 a fa a f fa ^h ^ • fa rH • fa fa 4 1 2 E o a

n n 12 fa ■g* • rH £ < c - s ~ H-H 4—» O cd 3 S 00 «* 4 n cd 166 2 i V cd fa N O N O 0 00 N O • H EC fa fa 3 fa fa - N > 4—4 4-4 4— 4-< 3 cd i 4 O N O 0 fa w —i y— -4 4 m m ’£ S . 3 r™H N ^t ^ (N I—H >—I .O fa > > fa 4 4-4 -4 4 4—> / C/5 o C/5 o i V i V d cd cd 0 2 cd E E O

cd f .0 fa • rH 1 H 4-> O •*—* O fa 0 2 3 c o >4

n 4 n c °u u g U 2 on f a f a fa^ fa^ a f a f fa m fa ^ m S ^ S e3 4 c 0 cd O O fa fa O 4 3 3 ~ - 3 - cd cd <-> § faO 3 C/5 • rH • C/5 2 3 fa a o 5 fa

m & • rH •rH W PQ > 4-» PH cd o cd O

CN o fa fa PQPQ U U fa U -3 on s s s C4 C4 O « ft H Cd »rH fa fa fa O O Q fa n +-4 i i -i— O G < C/5 00 G cj 2a G O «-* cd W> n n H ' 4—* b oo CX P CJ S o O 3 C o, O c a> O ON p*S u O S PQ P n h c c « Cx 3 s: §■ a s transplants (Ward et al., 1989, Wrzos et al., 1990, Okuno et al., 1990, Luka et al., 1990, Carrigan et al., 1991). Isolation of HHV6 from kidney biopsy (Asano et al., 1989, Okuno et al., 1990), bone marrow, sputum and broncheolar lavage material (Carrigan et al., 1991) from transplant recipients has also been reported.

Most of the isolates of HHV6 from adult peripheral blood have been obtained from patients with immune deficiency, either of iatrogeneic origin or due to infection with HIV and or other retroviruses. The exceptions to this are the isolates from CFS patients (Ablashi et al., 1988, Straus et al., 1988, Daugherty et al, 1991, Luka et al., 1990, Josephs et al., 1991) and the isolate obtained from an immunocompetent patient with pneumonitis (Russler et al., 1991). Many people have cultured peripheral blood from healthy adults on a regular basis but HHV6 isolates were not easily obtained until PBMC from patients with immune deficiency were cultured. In many cases the number of samples cultured before an isolate of HHV6 was obtained has not been reported. Downing and colleagues (1987) isolated HHV6 from only 2 of 500 PBMC samples from Ugandan AIDS patients and Becker et al. (1988) cultured PBMC from 60 HIV-1 infected patients, 9 HTLV-1 infected patients and 29 patients with lymphoproliferative disease without obtaining further isolates of the virus.

Yamanishi et al. (1988) first reported isolation of HHV6 from the peripheral blood of children during the acute phase of roseola infatum. Since this time many more HHV6 isolates have been obtained from PBMC of children with similar symptoms of rash and fever (Table 4. 2). Asano et al. (1989a, 1991) obtained HHV6 by cultivation of peripheral blood and plasma taken from children during the acute phase of roseola infantum. Thus the virus may be found cell free in peripheral blood during acute illness. Other HHV6 isolates have been obtained from PBMC of 2 infants with hepatitis (Asano et al., 1990a, Taijiri et al., 1990), a child with haemophagocytic syndrome (Huang et al., 1990) and pediatric bone marrow transplant recipients (Yoshikawa et a l, 1991, Asano et al., 1991a). Yoshikawa et al. (1991) also reported isolation of HHV6 from bone marrow samples from some of the 10 children who had received transplants.

4a. 1. ultrastructural characterization

In most cases, HHV6 isolates were initially identified as herpesviruses by EM of infected cells and tissue culture supernatants. Biberfeld and coworkers (Biberfeld et al., 1987, Ablashi et al., 1988) undertook detailed analysis of the virus particle. They reported that the nuclei of infected cells contained both 'empty" and 'full' nucleocapsids of 95 - 100 nm. The typical icosahedral structure of the virus capsid was revealed by negative staining. The complete enveloped virus particle was approximately 200 nm in diameter. This report also noted the prominent tegument around the capsids in the

168 Table 4.2 HHV6 isolation from children — C/3 •—1 /£ , , /£ O 0 £ 3 S ' o / O O C/3 j a C/3 • ♦ rH Q 4 0 H 4 < u * • ^4 <— cd CD C/3 J-h ^H 0 0 o c V S3 CJ 03 C h ( 03 < d i f f i f f 0 S 0 " S I 2 P £ v - r O d > C cd • • d 4 00 O 00 00 • • JH ^—4 t s PQ s - djd j cd • 1-H o * <+ 3 . t s • pH 4-» £ cd 3 V 4—» 3 S p a a cd p O oo D on G 3 (-< 3 3 s cd 4—» ) W cd 3 . t 0 C/i 0 ) 7 3 £ £3 3 pH ft 3 C/5 O W cd rH cd 0 I 1 n h n O OO G < r-H O (N ' S P S - O 3 _ <4-H o * 2 . '—*■ <4-1 4^» w 4-H0/2 4 C/5 cd <3 3 O O C/5 i_, 3 3 , a —• • 1 1 h n n , N O ON O ' ' O O r-l CO s 5 O NO OO P QQ 3 * 1 O H • rH • pH <+H <+H f s m O o c m U o * 2 , 2 - 2 . ' 2 . 169 — 4—» 4—» 4 3 O 3 P C/5 C/5 " —> h h n O < OO 4 ^ * t O , 3 H - C—

4-4 0 0 cd 3 O Q cd n n G 'o 4—1 1:3 00 00 3 £ 1 O 0 O O • ^4 Os s' O < O CN o ' • pH PQ s S s s s 2 . 1 H 4-* 4 4—» 4-4 O cd su C/5 H h 1 , O 0 O T"H t S ' P W Tp QPQ PQ • pH <4H 3 ' 2 . Si- H --- 4-4 4 4—1 <3 3 1) 3 e C/5 O 0 3 3 3 3 C 3 3 cx, cd 3 00 —) 1 1 n h n n

.3 13 3 r1'3 r ^ ■rs - ^H r-H 4—» . ^ c 00 ^ 0 O 3 cd O c £ £ 3 T 3 G ZT 0 3 C

c ?3 O "H 1 1 1 1 h h n 3 . «4-H ,3 ^ 4 0 1 —1 O O O 0 < H-4 QPQ ^ p PQ <4-H •£! • pH .cd x: *4—‘ C+H d I *4—* 4-» 4 - 4 4H C/5 cd C3 cd 3 O __ a C/3 cd 4> t f cd H p p I n n O O 0 m O a p ^ P s 8 • o ' • pH <4-H • O ■P 4—1 H-* cu C/5 NJ PH O (U 3 3 *H 3 -i—> o p cd ccP a 3 £ h x 2 X ,o3 Id O O U X XI "■ ffi O Pk PQ S -4-1 4—> o o o on o3 P a §• 3 bft 4 4— O a o a P 3 3 bi) 3 n n ch 4-1 • 3 g S • J2 3 •i-< T3 'o c ^ C/3 -4—4 S..2 3 I P « 1/3 S 03 c P c/3 O cd § * O 3 l-H c/3 P O CO O O u CM , a< PQ S .S —( bX) 3 ^3 3 n n cd C/5 170 O O O <+ X • >< ’o U PW PQ S rH O o C/3 cd 3 a 1-H P o3 3 P h n n h u s o X a & pq .&■ CM n i X • ^4 § * 4 3 3 3 p S e P P i-i cd 3 3 l-i P O O P 3 X £ . * — n O X £ O >< a X *g h o C/5 3 3 £ P n n

OO n o ’§u P 3 . ^ a 4} ao o a / VO V C/3 3 3 ^ 3 / V5 Q § vo 5 V 2 C/5 PQ ^ 6 3 ^ ^ p 3 g U O CM »0 h 3 Oh, C/3 rQ X a 5? Ph w cd cd n O O ^ 3 3 CM a C/3 cd o n n h h _ £ O O CM U 6 PQ 5 CM o _ <4—» 3 3 O P P n n h

CO CM .S ^ a . X 'H 4 * _ •4—* O P 2 g o cd cd 3 a^ fl COM ON ON CO O O PQ £ § 3 h 4 ^ h • rH ■s <-4-H P , ' t , zt -H 1-4 3 £ P >

Q PQ PQ - a 3- X p 3 ' 5 X ffi > p , d n -4—4 3 ° ^ 2 3 o o p p 3 _ ON ON u < O PQ s X 3 h cytoplasm of infected cells. Two other research groups confirmed this observation (Yoshida et al ., 1989a, Roffman et al, 1990). The latter group suggested that HHV6 first acquires the tegument in intranuclear compartments (tegusomes) and then acquires the virus envelope after budding into cytoplasmic vesicles.

4a. 2. cell tropism and In vitro growth characteristics

Salahuddin and colleagues (1986) originally reported that the newly isolated human herpesvirus (originally designated HBLV) infected only cells expressing B-cell-specific markers. Attempts to passage the virus in continuous cell lines of B- and T-cell origin were unsuccessful at this time. Later this group reported optimum culture conditions for the GS isolate (Lusso et al ., 1987, Lusso et al ., 1988). They confirmed that the majority of HHV6 infected cells in non-continuous cultures of adult peripheral blood, cord blood and thymus were of T-cell origin. The predominant phenotype of HHV6 infected cells in vitro was CD4, CD7, CD2 positive. Infected cells gave only weak cytoplasmic staining with an anti-CD3 MAb and lacked the receptor for IL-2. Some of the infected cells coexpressed CD4 and CD8. Thus most of the cells infected with the GS virus in these cultures were immature T-lymphocytes.

Supernatant virus was produced in GS infected adult PBMC and filtered medium from infected cells could be used to passage the virus. Culture of the GS virus in continuous cell lines of B- and T-lymphocytes, megakaryocytes and glial cells was then reported (Ablashi et al ., 1987). Infection of the immature T-cell line HSB-2 (CD3 and CD4 negative) and the mature T-cell line JM 2.7 (a CD4 positive, CD8 negative clonal derivative of Jurkat) was also reported to produce large quantities of supernatant vims (Ablashi eta l ., 1988b).

The U1102 isolate of HHV6 was originally propagated by passage of infected cells into PHA and IL-2 stimulated cord blood cells, very little if any supernatant vims was produced. The vims was reported to replicate in cell lines of T-lymphocytic, B- lymphocytic, and monocyte / macrophage lineage again with the production of very little supernatant virus.

The AJ isolate of HHV6 was obtained by cocultivation of the patient's PBMC with PHA stimulated cord blood cells. The vims was subsequently passaged into T- lymphocytic, B-lymphocytic, embryonic lung and embryonic glial cell lines (Tedder et al ., 1987). Supernatant vims was initially detected in cultures of AJ infected cord blood cells but not when passaged into the J Jhan T-cell line (a CD4 positive clonal derivative of Jurkat).

171 During the course of this study, the AJ virus was passaged from J Jhan cells into other T-cell lines (including HSB-2) and back into PHA stimulated cord blood cells but there was no detectable supernatant virus produced using the culture methods and conditions described. The majority of research groups who regularly culture the AJ and U1102 isolates of HHV6 report that there is no supernatant virus produced in continuous cell lines. Two laboratories have been able to purify AJ from the tissue culture supernatant of infected J Jhan (Dr P. Coyle, personal communication) or HSB-2 (Dr H. Kangro, personal communication) cell lines. The reason for this discrepancy is not clear but may be due to adaptation of the virus or differences in growth conditions. Throughout the course of the study reported in this thesis the AJ isolate of HHV6 propagated in the J Jhan cell line was used as source of viral antigen.

The fourth isolate of HHV6 which was reported in the literature and extensively characterized was also obtained from PBMC samples of an HIV infected individual (Lopez et al., 1988). This virus (Z29) was propagated in cord blood cells with the production of very little supernatant virus. Again, the majority of infected cells were of T-lymphocytic origin but attempts to passage the vims into continuous cell lines were largely unsuccessful. Only one cell line of T-cell origin (HTLV-1 infected MT-4 cells) was found to support the replication of Z29 (Black et al., 1989) and this was very limited. The authors recommended the propagation of Z29 in cord blood cells stimulated with PHA, IL-2 and hydrocortisone to enhance the production of supernatant vims.

An isolate of HHV6 obtained from the saliva of an HIV infected individual (HHV6Sp) has been characterized (Levy et al., 1990). This isolate was found to preferentially infect CD4 positive T-lymphocytes. Adult PBMC were found to be better than stimulated cord blood cells for the propagation of this isolate with the production of supernatant vims. The only T-cell line found to adequately support the growth of HHV6SF was the MT-4 HTLV-1 infected continuous cell line. The vims did not replicate in the HSB-2 immature T-cell line. Low but persistent levels of vims replication were noted in freshly isolated human macrophages and cell lines derived from colorectal and neuroblastoma cells.

Wyatt et al. (1990) compared the replication of Z29 and U1102 isolates of HHV6 in PBMC and continuous cell lines. They confirmed that U1102 but not Z29 would productively infect J Jhan and HSB-2 continuous cell lines using standard medium constituents. When PHA was added to J Jhan cultures and these then infected with Z29 a small number of cells expressed HHV6 antigen. In addition, U1102 would replicate in PBMC which had not been stimulated with PHA but Z29 required T-cell activation for productive infection.

172 Takahashi and colleagues (1989) reported that HHV6 was isolated from CD4 but not CD8 positive peripheral blood mature T-lymphocytes of patients with roseola infantum. In vitro studies confirmed that these HHV6 isolates selectively infected CD4 positive mature T-lymphocytes which expressed the IL-2 receptor. The Hashimoto isolate of HHV6 (Yamanishi et al., 1988) was found to replicate in freshly isolated adherent monocytes which had been cultured for 7 days prior to infection. When these monocytes were cultured for 1 month prior to infection (and had presumably matured to macrophages) the virus did not seem to replicate but infectious virus was produced when the cultures were treated with phorbol ester (Kondo et al., 1991). The authors suggested that HHV6 may latently infect these cells. Seven T-cell lines were tested for susceptibility to infection with the Hashimoto isolate of HHV6. The MT-4 cell line was found to give the highest yields of virus (intra and extracellular) and was used to prepare an HHV6 neutralization assay (Asada et al., 1989).

Kikuta et al. (1990) reported that the OK isolate of HHV6 obtained from a roseola infantum patient (Kikuta et al., 1989) productively infected cord blood cells but not adult PBMC. They found that addition of a MAb to CD3 enhanced viral replication in adult PBMC presumably by boosting T-cell proliferation. These results confirmed that the HHV6 isolates from roseola infatum patients predominantly infect mature T- lymphocytes as previously reported (Takahashi et al., 1989).

All isolates of HHV6 which have been characterized replicate preferentially in cells of T-lymphocytic origin. Variations in the ability to propagate isolates of HHV6 in continuous cell lines, with or without the production of supernatant virus, may be due to laboratory adaption of the virus or cell line or genuine differences in growth characteristics of these viruses. Luka et al. (1990) used the MRC-5 cell line (derived from human lung fibroblasts) for isolation and propagation of HHV6.

Ablashi and colleagues (Ablashi et al., 1988, Ablashi et al., 1988b, Straus, 1988) reported that the two isolates of HHV6 obtained from adults with CFS did not replicate in the HSB-2 cell line as well as the original GS isolate. Takahashi and coworkers suggested that discrepancies in the cell tropism of HHV6 when comparing isolates obtained from roseola infantum (Takahashi et al., 1989) and HIV infected individuals (Lusso et al., 1987, 1988) may be due to differences in cell culture conditions. Addition of PHA to cultures may induce expression of CD3 and IL-2 receptor. In purified T-cell cultures, ie. in the absence of accessory cells, IL-2 will not be produced and T-cell proliferation will not occur unless exogenous IL-2 is added to the culture. If unfractionated peripheral blood cells are used for virus propagation it may not be necessary to add IL-2 for T-cell proliferation (Frenkel et al., 1990a). The Z29 isolate of HHV6 required T-cell activation for growth but Roffman and Frenkel (1990) reported

173 that if too high a concentration of IL-2 was added to PBMC cultures the virus CPE was inhibited. The mechanism involved in this inhibition and the implication for in vivo virus infection is unknown.

The in vitro host range of HHV6 may not be restricted to cells of human origin. Two groups reported that freshly isolated PBMC from chimpanzees could support replication of the virus (Levy et al, 1990, Lusso et al ., 1990).

4a. 3. sequence variation amongst HHV6 isolates

Although different isolates of HHV6 were shown to be related by molecular hybridization studies, restriction enzyme pleomorphism amongst these isolates was reported by a number of research groups (Josephs et al ., 1988, Becker et al., 1989, Jarrett et al., 1989, Kikuta et al. , 1989, Levy et al., 1990, Schirmeret al., 1991, Aubin et al., 1991). These differences may account for discrepancies in the reported cellular host range for these viruses.

Schirmeret al. ( 1991) reported that the restriction enzyme pattern for the Z29 isolate of HHV6 was very similar to that for HHV6 isolates obtained from patients with roseola infantum. The only differences in restriction enzyme patterns for these isolates was within the terminal repeat sequences. When restriction enzyme analysis of the U1102 virus was carried out, the pattern was very different to that of the Z29 and roseola infantum isolates. The authors suggested that HHV6 isolates may be divided into two groups based on their restriction enzyme patterns. All available information indicated that the two other original isolates of HHV6 (GS and AJ) were most closely related to the U1102 virus when analyzed with restriction enzymes and were distinct from the Z29 and roseola infantum isolates (Ablashi et al., 1991, Dr R. Honess and Dr B. Thomson, personal communication).

Aubin et al. (1991) studied the genetic polymorphism of 8 different HHV6 isolates using PCR and restriction enzyme analysis. They reported that these viruses could also be divided into two groups based on restriction enzyme patterns. These groups were typified by the SEE (Agut et al., 1989, from an HIV-2 infected adult) and HST (Yamanishi etal. , 1988, from roseola infantum patient) isolates of HHV6.

4a. 4. antigenic differences between HHV6 isolates

Initial characterization of HHV6 isolates confirmed a lack of antigenic crossreaction with the other human herpesviruses (Salahuddin et al., 1986, Tedder et al ., 1987, Downing et al., 1987). Since that time, however a MAb which neutralizes EBV, HHV6, CMV and bacteriophage T4 DNA polymerases has been reported (Tsai et al., 1990)

174 suggesting that these polymerases have a common epitope.

Balachandran and colleagues (1989, 1991) identified at least 33 proteins which were specific for HHV6 infected tissue culture cells. Other research groups (Shiraki et al., 1989, Okuno et al ., 1990a, Yamamoto et al., 1990) identified 20 - 29 HHV6 virion- associated proteins.

The antigenic relationship of different HHV6 isolates has been investigated by a number of research groups. One group (Wyatt et al., 1990) used 6 MAbs which were prepared against the GS isolate of HHV6 to investigate antigenic differences between Z29 and U1102 infected cells. Each monoclonal had previously been shown to react with a distinct protein found in GS infected tissue culture cells (Balachandran et al ., 1989). All MAbs reacted with cells infected with the U1102 isolate but only 4 reacted with cells infected with Z29. These results confirmed that the U1102 and GS isolates of HHV6 were more closely related to each other than to the Z29 isolate. These results were extended by Schirmer et al. (1991) who reported that cells infected with 7 isolates of HHV6 from patients with roseola infantum resembled Z29 and not U1102 infected cells in their reaction with MAbs. Antigenic variation has also been noted when comparing the virion polypeptides and infected cell associated polypeptides for different roseola infatum isolates of HHV6 (Yoshida et al., 1989, 1990).

These results suggest antigenic variation amongst different isolates of HHV6 and also variation in the individuals antibody response to the virus.

4b. Detection of anti-HHV6 in healthy individuals

In this study, serological assays for the detection of anti-HHV6 were prepared and used to determine the seroprevalence of HHV6 in blood donors. Indirect and ACIF were developed for detection of anti-HHV6 IgG and sera from blood donors were screened for HHV6 antibody using these assays and scored acording to the intensity of the fluorescent staining. A proportion of sera screened for HHV6 antibody by indirect IF gave distinct, but dull, staining of the expected number of cells. These sera were designated equivocal in this assay and could not be resolved by ACIF. For antibody prevalence studies these sera were classified as negative. There was no difference in the sensitivity of the two IF assays and indirect IF was used for anti-HHV6 seroprevalence studies. Using this assay, approximately 60% of blood donors were classified as anti- HHV6 IgG positive and the number of strongly reactive sera was significantly higher in female than in male blood donors.

A competitive RIA was developed to try and evaluate sera that were weakly reactive by IF. Sera were tested undiluted in this assay and the % inhibition of the label bound

175 calculated. The histogram for the blood donor group was bimodal with a small proportion of sera competing very efficiently in this assay. It is possible that all sera from blood donors were anti-HHV6 positive by RIA but that some had only very low levels of antibody. All of the blood donor sera tested gave some inhibition of label binding when compared with FCS. Serum samples from young children prior to HHV6 primary infection and documented seroconversion (by indirect IF) also gave some inhibition of label binding in the RIA. These samples represented the closest to true anti-HHV6 negative sera available although they may have still contained low levels of maternal antibody. Based on these results a scoring system for anti-HHV6 by RIA was developed and used for the estimation of antibody prevalence in male and female blood donors. By RIA, 80 % of sera from blood donors were anti-HHV6 positive and there was no significant difference in the number of strongly reactive sera when comparing males and females in this assay.

The correlation between anti-HHV6 by IF and RIA for the group of blood donors was good, many of the sera designated equivocal by IF were clearly positive by RIA. The group of blood donor sera were also tested for antibody to the other human herpesviruses. There was no correlation between presence or level of HHV6 antibody by IF or RIA and presence or level of antibody to any of the other human herpesviruses. These results confirmed the high specificity of IF and RIA for the detection of HHV6 antibody. There was no correlation between age of blood donor and anti-HHV6 presence or titre by IF or RIA.

A number of other research groups have developed serological assays for the detection of anti-HHV6. Indirect and ACIF have been most extensively used but the seroprevalence rates quoted for anti-HHV6 by these assays have varied greatly. Robert et al. (1990) confirmed that there was no difference in sensitivity when comparing indirect and ACIF for detection of anti-HHV6. Differences in the initial dilution for antibody screening and the subjective nature of IF make variation in results between laboratories almost inevitable. It is possible that antigenic differences between HHV6 isolates may account for some of the discrepancies in antibody prevalence reported. Enders et al. (1990) obtained discrepant results when different isolates of HHV6 were used to detect specific antibody by indirect IF. Yadav et al. (1991) reported differences in antibody titre when screening sera by indirect IF using the GS and Z29 isolates of HHV6. Antigenic differences between HHV6 isolates were not thought to be the main source of these variations, however since many of the laboratories reporting discrepant anti-HHV6 results were using the same HHV6 isolate as source of antigen. It was noted during the course of this study that the length of time that the cells had been infected before the preparation of IF slides was important in the reproducibility of this assay.

176 Most people now agree that a large proportion of sera from adults are anti-HHV6 positive and that specific antibody levels are generally low in the adult population. One group confirmed the observation that the number of strongly reactive sera by indirect IF was higher in female than in male blood donors (Clark et al., 1990). When anti-HHV6 was measured by ELISA (Saxinger et al., 1988, Asano et al., 1990) there was no difference in antibody prevalence or titre when comparing male and female blood donors. Okuno et al. (1989) reported that the presence and level of HHV6 antibody in adults did not vary with age. Other research groups (Brown et al., 1988, Levy et al., 1990a, Yanagi et al., 1990), however noted a decline in anti-HHV6 IgG prevalence and titre with increasing age. These discrepant results are probably due to differences in the sensitivity of these assays for the detection of anti-HHV6.

Two research groups have reported the use of anti-HHV6 IgG ELISAs for the determination of HHV6 seroprevalence in blood donors (Saxinger et al., 1988, Dahl et al., 1990). Saxinger and colleagues (1988) reported that the ELISA test results were normally distributed with no evidence of bimodality and no obvious cut-off between positive and negative. In a similar way to that described in this study, Saxinger et al. (1988) compared anti-HHV6 with antibody to the other human herpesviruses in this blood donor population and confirmed the specificity of their assay. Linde et al. (1990) found no evidence of bimodality in sera from 40 healthy adults tested for anti-HHV6 IgG by ELISA. The cut off for this assay was fixed at the ELISA OD for 10 IF negative sera ± 2 SD. Approximately 90 % of these adult sera were HHV6 antibody positive using this criteria for assessment of positive and negative.

The apparent bimodality of the anti-HHV6 RIA histogram, which was not reported for anti-globulin ELISAs, may be due to discrimination between high and low avidity antibody by the former method. Specific antibody of the IgM, IgA, IgD and IgE class may also compete with the anti-HHV6 radiolabel in the RIA. These antibodies would not be detected by the methods described for anti-HHV6 ELISA (Saxinger et al., 1988, Dahl e ta l, 1990).

Other serological assays reported for the detection of anti-HHV6 include western blot, RIPA and virus neutralization assays. Western blot and RIPA assays have not, as yet, been widely used for the detection of anti-HHV6 in healthy individuals. Where studies using these assays have been undertaken, sera from most adults were reported to give only weak reactions with HHV6 specific polypeptides (Saxinger et al., 1988, Ablashi et al., 1988) although the reaction with a putative nucleocapsid protein was strong (Yamamoto et al., 1990). This may be due to the low levels of HHV6 antibody in most adults' sera or problems in the preparation of immunologic ally reactive polypeptides. Other groups have used RIPA for the identification of patient groups with high titre

177 anti-HHV6 (Balachandran et al. 1989, Josephs et al., 1991). A neutralization assay for anti-HHV6 has also been developed and used to confirm HHV6 primary infection of children (Asada et al., 1989, Asano et al., 1989a, Yoshida et al., 1989, Yoshida et al., 1990, Yoshikawa etal., 1990, Ablashietal., 1991).

Seroepidemiological studies of HHV6 in adults from different geographical areas have been undertaken (Ablashi et al., 1988, Levine et al., 1990, Giraldo et al., 1990, Rodier et al., 1990, Yadav et al., 1990, Yadav et al. 1991, Ranger et al., 1991). These studies confirmed the worldwide distribution of HHV6 but indicated that the prevalence and titre of anti-HHV6 may vary across different continents.

The indirect IF assay was used to screen sera from children aged < 1 year - 17 years for anti-HHV6 IgG. Maternal antibody was demonstrable in children up to 6 months of age. After this time there was a rapid acquisition of anti-HHV6 until by 1 year of age the HHV6 antibody prevalence was similar to that found in older children and not significantly different to that found in adults. It was concluded from these results that the majority of individuals acquire HHV6 very early in life. This early acquisition of HHV6 was also reported by a number of other research groups (Yamanishi et al., 1988, Takahashiet al., 1988, Brown et al., 1988, Pietroboni et al., 1988a, Knowles and Gardner, 1988, Saxinger et al., 1988, Okuno et al., 1989, Yoshikawa et al., 1989, Ueda et al., 1989, Enders et al., 1990, Farr et al., 1990). Farr and colleagues (1990) confirmed that acquisition of HHV6 generally occurred earlier in life than acquisition of CMV, HSV, VZV and EBV.

An indirect IF test for the detection of anti-HHV6 IgM was developed. None of a group of blood donor sera were HHV6 antibody positive by indirect IF and this assay proved useful for the identification of recent HHY6 infection (see 4d. and 4e.). Attempts to prepare an IgM capture RIA for the detection of anti-HHV6 IgM based on the technology used for anti-VZV (Tedder et al., 1981) proved unsuccessful. If supernatant virus had been available as source of antigen this may have been possible but the use of cell lysates prepared from AJ infected J Jhan cells gave very high backgrounds in a non competitive format assay. One research group has reported use of an ELISA for detection of HHV6 IgM antibody (Saxinger et al., 1988, Steeper et al., 1990). The authors reported an increase in HHV6 IgM antibody up to 1 year of age but gave no details of specificity of this assay.

4c. Tissue localization of HHV6

Methods for the detection of HHV6-specific DNA and protein in tissues and body fluids were developed. These methods were used to identify potential sites of HHV6 persistence and possible routes of transmission of the virus.

178 Nested PCR proved to be a very sensitive technique for the detection of HHV6 DNA and great care had to be taken to avoid contamination of the reaction mixture. When plasmid controls diluted in negative DNA were amplified by this method < 5 copies of HHV6 sequence was detectable after two rounds of PCR.

Many HHV6 isolates were obtained by culture of Ficoll-Paque separated PBMC from adult patients. Only two groups, however have obtained HHV6 isolates from PBMC of healthy adults (Lopez et al., 1988, Daugherty et al ., 1991). In order to investigate this, peripheral blood samples from healthy adults were extracted and screened for HHV6- specific DNA by nested PCR. Only a small number of circulating lymphocytes contained HHV6 sequence since the detection of specific DNA was dependant on the amount of extracted PBMC sample added to the reaction. Approximately half of the PBMC samples were HHV6 positive by nested PCR when the DNA equivalent of 105 cells was added to the PCR reaction mixture. These results are comparable with those reported by two other research groups (Jarrett et al., 1990, Gopal et al., 1990). Jarrett et al. (1990) reported that some PBMC samples may inhibit the PCR reaction. A control plasmid containing HHV6 DNA was diluted in 2 pg of the negative DNAs to confirm that the nucleic acid extracts used in this study were not inhibitory to the PCR reaction.

Kondo and colleagues (1991) infected adherent mononuclear cells (monocytes) in vitro with HHV6. No virus particles were produced after 35 days of culture but viral DNA was detectable by PCR. Infectious virus particles were produced when the culture was treated with phorbol ester. The authors suggested that HHV6 may latently infect macrophage (mature, differentiated monocytes) in vivo.

The results of HHV6 PCR studies on PBMC from healthy adults are similar to those reported for EBV and CMV. The detection of EBV and CMV DNA in circulating cells does not seem to be related to specific antibody levels in that individual but may be indicative of recent infection. Gopal et al. (1990) reported that approximately half of the PBMC samples from healthy adults were positive for EBV DNA by PCR, this was similar to the prevalence of HHV6 DNA in these individuals. Other research groups have not been able to detect EBV DNA in peripheral blood of healthy adults (Crouse et al., 1990, Telenti et al., 1990). One research group has reported detection of CMV DNA in the peripheral blood of anti-CMV positive and negative blood donors (Morris et al., 1989) but others have not been able to show this (Shibata et al., 1988, Jiwa et al, 1989). Discrepancies in the reported PCR results may be accounted for by inhibitors of amplification in some PBMC samples and differences in the amount of extracted DNA or whole blood added to the PCR reaction. The consensus of opinion is that HHV6, CMV and EBV sequences are present in only a small number of circulating mononuclear cells of healthy adults. An increase in CMV, or EBV DNA in peripheral

179 blood may be indicative of a recent herpesvirus infection (Shibata et al ., 1988, Crouse et al., 1990, Telenti et al., 1990). Amplification of HHV6 sequences from peripheral blood has been used to try and confirm primary infection and reactivation of HHV6 in patient groups (Buchbinder et al., 1988, Asano et al ., 1990a, Wrzos et al ., 1990, Gopal et al ., 1990, Kondo et al., 1990, Kondo et al., 1991).

Histologically normal submandibular salivary gland tissues and TW samples from healthy adults were extracted and screened for HHV6 DNA by nested PCR. All salivary gland extracts and half of the TW samples were positive for HHV6 DNA after two rounds of PCR. Salivary gland and TW samples were found to inhibit the PCR amplification when > 0.5 (ig of DNA was used in the reaction. Plasmid containing the HHV6 sequence was diluted in PCR negative TW DNA extracts to identify any inhibitory samples. The problem of inhibitory reactions was alleviated when < 0.5 fig of DNA from TW samples was used in the PCR reaction. Sample inhibition of PCR has been reported by a number of research workers (discussed in Higuchi, 1989), particularly when crude tissue or cell extracts are used as source of DNA. One research group reported that gel filtration and boiling removes the inhibitor of Taq polymerase (de Franchis et al., 1988). Unfortunately this method is time-consuming and could not be used when only small quantities of sample DNA are available.

None of the TW samples from healthy adults screened for HHV6 DNA were positive after a single round of PCR. As 50 copies of HHV6 plasmid were routinely detectable after one round of PCR it could be concluded that < 50 copies of HHV6 sequence were present in the 0.5 fig of DNA added to the PCR reaction. All HHV6 DNA positive samples amplified 8-globin sequences efficiently (maximum strength band after 1 round of PCR) suggesting that the HHV6-specific signal was not 'damped down' by low levels of inhibitor in the sample. One apparently HHV6 antibody negative individual was found to shed HHV6 into the throat, others were found to have intermittently detectable virus in TW samples.

The presence of HHV6 in oropharyngeal samples from healthy adults by PCR was reported by three research groups. Gopal and colleagues (1990) screened 30 extracted TW samples from healthy individuals for HHV6 DNA by PCR with concordant results to those reported in this study. Jarrett et al. (1990) amplified boiled saliva directly and reported that 90 % of healthy adults shed virus in saliva. Kido et al. (1990) found only 3 % of throat swab samples from healthy adults had detectable HHV6 DNA by PCR. Discrepancies between reported HHV6 PCR results are probably due to differences in the amount of DNA added to the reaction mix. Jarrett et al. (1990) also confirmed that some anti-HHV6 negative adults shed HHV6 into the throat and some anti-HHV6 positive individuals had intermittently detectable virus in oropharyngeal samples.

180 Two research groups (Pietroboni et al., 1988a, Harnett et a l, 1990, Levy et al., 1990a) reported isolation of HHV6 from saliva samples of most healthy adults but others have not been able to repeat these results (Yoshiyama et al, 1990, Lopez and Honess, 1990, Kido et aL, 1990). Most of these isolation studies utilised adult PBMC for culture of saliva samples. In view of the presence of HHV6 sequence in some PBMC samples, the isolation of virus from saliva must be interpreted with caution. Jarrett et al. (1990) postulated that saliva may 'reactivate" latent HHV6 in peripheral blood.

A biotinylated HHV6 probe was prepared and used for in situ localization of virus- specific DNA in tissue sections. As molecular studies in other laboratories had shown sequence homology between HHV6 and CMV (see 4i.) tissues were also screened for CMV-specific DNA. The HHV6 probe did not stain CMV positive tissues from an AIDS patient and the CMV probe did not react with HHV6 infected J Jhan cells. It was concluded that the in situ hybridization method was specific for the detection of HHV6 DNA.

Histologically normal submandibular and parotid salivary glands were found to contain high levels of HHV6 DNA by in situ hybridization. The viral DNA was localized mainly in the nuclei of mucous and serous cells although weak cytoplasmic staining of some cells was also noted. The striated ducts, which were most obvious in sections of parotid salivary glands, were also stained for HHV6 DNA. Although most of the submandibular glands also contained CMV DNA the proportion of mucous cells stained was less than stained for HHV6 DNA and staining of serous and columnar cells in submandibular glands was either absent or very weak. There was no detectable CMV DNA in parotid glands and it was concluded that both viruses were present independently in salivary gland tissues.

In order to determine if HHV6 protein was present in salivary glands, HHV6 MAbs were prepared and used to localize virus-specific proteins in tissue sections. The anti- HHV6 monoclonals did not react with uninfected J Jhan cells nor with the CMV infected control tissues. Tissues were also screeed with a CMV MAb, this monoclonal did not react with HHV6 infected J Jhan cells. The presence of HHV6-specific proteins was confirmed in 8 of the 9 submandibular glands and both of the parotid glands that contained HHV6-specific DNA. The HHV6 protein was mainly localized in the cytoplasm of mucous cells, the nuclei of serous cells and both nucleus and cytoplasm of columnar cells. Antigen specific for CMV was shown in all the salivary glands that were positive for CMV DNA but not those that were negative. Serous and columnar cells did not stain for CMV antigen.

181 From all these results it would seem that many adults shed HHV6 into the oropharynx and that the source of this virus may be persistently infected salivary gland tissue. One tissue sample was found to contain detectable HHV6 DNA but not protein using the methods described. All other tissues gave concordant results by in situ hybridization and immunohistochemical staining. As the anti-HHV6 monoclonals had not been characterized it was not possible to distinguish between latent and productive infection of salivary gland tissue. It would seem likely that infected salivary gland tissue is the source of cell free virus in saliva. If a MAb prepared against 6 or y HHV6 proteins was available it would be possible to confirm that HHV6 infection of salivary gland tissue is productive.

Krueger and colleagues (1990) confirmed the observation that HHV6 DNA and protein is present in salivary gland tissue. They reported that 9 of 12 lip salivary glands taken as a biopsy from healthy individuals were HHV6 positive by immunohistochemistry. The MAb used (9A5D12) was specific for the p41 protein of HHV6 infected cells (Balachandran et al, 1989). This protein has been partially characterized (Chang and Balachandran, 1991) and although it is detectable early in infection, DNA replication is required for continued synthesis. Detection of large quantities of this protein within salivary gland tissue probably indicates productive HHV6 infection.

Only nuclear staining of serous acini for HHV6 DNA and protein was demonstrable by in situ hybridization and immunohistochemistry. It is possible that HHV6 latently infects serous cells and productively infects mucous and columnar cells. This study was the first to identify salivary gland tissue as a site of HHV6 persistence and a possible site of HHV6 latency. The demonstration of cell free virus in saliva was concordant with published isolation studies and suggested an air-boume route of transmission for the virus.

Cytomegalovirus-specific DNA and protein was demonstrated in some mucous acini and striated ducts. Interactions between HHV6 and CMV may be important in the establishment of latency, productive infection and transmission of these viruses.

Lymph node ( Eizuru et al., 1989, Borisch Chappuis et al, 1989, see 4e. 3. and 4e. 5.) large bowel tissue (4e. 9.) and kidney (4e. 6.) have also been identified as potential sites of HHV6 persistence in healthy individuals. Russler et a l (1991) reported detection of HHV6 antigen, by immunohistochemical staining, in lung biopsy material from an immunocompetent patient with pneumonitis and concomitant Legionella infection.

182 4d. HHV6 in pregnancy - implications for transmission

The early acquisition of HHV6 suggested that infection of the newborn infant commonly occurs during the perinatal period. Initial serological studies established that the prevalence of HHV6 IgM antibody increased during pregnancy but none of a group of cord blood sera were anti-HHV6 IgM positive. Farr and colleagues (1990) also screened cord blood samples for HHV6 specific IgM antibody, with negative results. Harnett et al. (1990) were unsuccessful in their attempts to culture HHV6 from endocervical swabs. All these results indicate that neonatal HHV6 infection (either in utero or at birth) is not a common occurrence. Takahashi and colleagues (1988) compared the incidence of HHV6 primary infection of infants who were breast fed with those who were bottle fed and concluded that breast milk is not a common route of transmission of the virus. Horizontal transmission of HHV6 from adults to young children seemed likely.

A prospective study of HHV6 in pregnancy and the early postnatal period was undertaken to confirm serological reactivation of the virus and to determine the presence and level of HHV6 DNA in saliva and peripheral blood by nested PCR in these women. Serological markers for reactivation of HHV6 were defined, for the purposes of this study, as the detection of specific IgM antibody in one or more serum samples and or a significant rise in anti-HHV6 IgG titre in a pregnant woman. The majority of pregnant women enrolled in this study had detectable HHV6 IgM antibody in one or more sera with or without significant changes in anti-HHV6 by IF or RIA. In many women PBMC samples taken during the course of this study fluctuated between HHV6 PCR positive and negative, presumably reflecting the low numbers of circulating cells containing HHV6 DNA. The number of women shedding HHV6 in saliva and the amount of virus shed in saliva (as determined by the number of first round PCR positives) increased with advancing gestation. Approximately 90 % of the 70 women in this study had detectable HHV6 DNA in TW taken late in pregnancy or during the early postnatal period, this was significantly more than shed virus during the first trimester. Many of the third trimester and early postnatal TW samples were confirmed as HHV6 DNA positive after a single round of PCR, a phenomenon that was not seen when TW from non-pregnant adults were screened for HHV6 DNA.

These results confirmed that serological reactivation of HHV6 occurs in at least some pregnant women and that the amount of virus shed in the oropharynx is boosted late in pregnancy and early postpartum. The high levels of HHV6 shed by mothers during the postnatal period may result in transmission of the virus to their newborn offspring. The early acquisition of HHV6 indicates that infection may occur in the presence of maternal antibody.

183 Cytomegalovirus is one of the most common viral infections to be transmitted in utero. Congenital CMV infection may occur as a result of primary infection or reactivation in the mother and can also occur in the presence of high levels of anti-CMV (Reynolds et al ., 1973). Cellular immunity plays a part in prevention of intrauterine transmission of CMV (Rola-Peszczynski et al., 1977, Gehrz et al., 1977, Starr et al., 1979, Reynolds et al., 1979, Stem et al., 1986) and this may be important also in the case of HHV6 transmission. In contrast to CMV, there was no evidence of congenital HHV6 infection, either from this study or from other published reports.

Both CMV and HSV may be acquired at birth by contact with infected genital secretions. An increase in cervical and urinary shedding of CMV with advancing gestation was reported by two research groups (Reynolds et al., 1973, Stagno et al ., 1975). In both studies CMV excretion began late in pregnancy and continued into the postnatal period with no evidence of CMV viraemia or changes in CMV antibody profile in these women. The authors suggested that the shedding of CMV without evidence of viraemia may be a result of localized reactivation and replication of the virus. It is possible that localized reactivation of CMV and HHV6 occurs as a result of hormonal stimuli and allows transmission of the virus to the infant during the neonatal or perinatal period. Many more women shed HHV6 in saliva than shed CMV in cervical secretions, breast milk, urine or saliva during late pregnancy and the early postnatal period. This may account for the greater number of HHV6 than CMV infections that occur during the first year of life.

4e. HHV6 potential disease associations

Many research groups have used HHV6 antibody and or virus detection to investigate clinical symptoms and diseases which may be associated with HHV6 primary infection and reactivation. Serological studies of healthy adults confirmed the ubiquitous nature of HHV6. The association of such a ubiquitous virus with the pathogenesis of disease is difficult to prove.

4e. 1. roseola infantum

Roseola infantum (exanthem subitum or sixth disease) is a common febrile illness of early childhood that was first described over a century ago. It was first suggested in the 1950s that roseola infantum was caused by a ubiquitous virus to which adults and older children were immune. Transmission of the disease inducing agent between infants and from infants to monkeys was demonstrated by inoculation of filtered serum subcutaneously or throat washings sprayed intranasally (reviewed by Hall, 1989). Early studies demonstrated that males and females suffered from roseola infantum equally

184 and that the majority of cases occurred in children under 3 years of age (Berenberg et a l, 1949).

Roseola infantum is characterized by a constant or intermittent fever that lasts for 3-5 days, the temperature usually drops to normal with the appearance of the classical roseola rash. The rash occurs predominantly on the neck and trunk but may spread to involve other parts of the body. The eruption is characterized by discrete, irregular macules. The eruption is often transient and may be visible for only a few hours or up to 2 days. Pharyngitis and cervical lymphadenopathy are the most common secondary symptoms associated with roseola infantum (Stammers, 1988).

Yamanishi and colleagues (1988) first identified HHV6 as the causative agent of roseola infantum in young children. They isolated HHV6 from the peripheral blood of 4 infants with clinically diagnosed roseola infantum and demonstrated seroconversion for anti-HHV6 IgG by ACIF in these patients. The isolation of HHV6 from peripheral blood of roseola infantum patients was later confirmed by other research workers (Asano et al., 1989a, Kikuta et al, 1989, Yoshida et al, 1990, Enders et a l, 1990, Portolani et a l, 1990, Yoshiyama et al, 1990). Asano et a l (1989a, 1991) reported isolation of cell free virus from plasma during the acute phase of roseola infantum. After 8 days of illness it was not possible to isolate HHV6 from periperal blood, this coincided with the detection of anti-HHV6 neutralizing antibody in these patients. Yoshida et a l (1989) reported that the HHV6 antibody profile was different for infants with roseola infantum than for their mothers. These results confirmed that the HHV6- specific antibody detected in sera from these children was not maternally derived.

In the study reported in this thesis, sera from children with clinically diagnosed roseola infantum or other febrile illnesses of childhood were screened for anti-HHV6 IgM and IgG by indirect IF and competitive RIA. Anti-HHV6 specific IgM was detected in at least one serum sample from 8 of the 9 children with clinically diagnosed roseola infantum (serum from the other child gave equivocal results in this assay). Seroconversion for anti-HHV6 IgG, by indirect IF, was demonstrated in the other samples from 8 children. A number of other research groups also demonstrated HHV6 specific IgM antibody in sera from children with roseola infantum (Knowles and Gardner, 1988, Enders et al, 1990, Kondo et al, 1990, Irving et a l, 1990a, Asano et al, 1991b).

To confirm the specificity of the anti-HHV6 IgM indirect IF assay, all sera that were positive for HHV6 IgM antibody were also screened for anti-CMV IgM, all were negative. Acute and convalescent samples from children with rubella or enterovirus infections were also screened for anti-HHV6 IgM, with negative results. Another

185 research group (Enders et al., 1990) also tested the specificity of indirect IF for the detection of anti-HHV6 IgM and found that children with acute viral infections such as rubella, measles, mumps and varicella were all negative for HHV6 IgM antibody.

Many more children seroconvert for anti-HHV6 than are accounted for by cases of clinically diagnosed roseola infantum. Primary HHV6 infection of most young children probably produces only very mild clinical symptoms, or no symptoms at all. In this study 2 cases of anti-HHV6 seroconversion that were not associated with 'classical' roseola infantum were identified. One child had pyrexia but no roseola rash, similar to the patient described by Suga et al. (1989). It is possible that in these 2 cases the rash was very transient and went unnoticed. Asano and coworkers (1989b) reported HHV6 primary infection associated with rash, but no fever in 2 infants. In the study reported in this thesis 1 child who seroconverted for anti-HHV6 suffered from pyrexia and rash but also had concomitant gastroenteritis. A concurrent rotavirus infection may have been the cause of the gastric symptoms in this patient but an exacerbation of this by the HHV6 infection could not be ruled out. As HHV6 infection in childhood seems to be very common, concomitant infection of an individual with HHV6 and other viruses may occur frequently. Suga and coworkers (1990) reported simultaneous infection of 2 infants with HHV6 and measles (4 - 5 months of age) which manifested as an atypical roseola infantum / measles-like illness. There has also been a report of HHV6 reactivation or reinfection associated with a recurrence of clinical symptoms in some children (Kusuhara et al., 1991).

Some infants may suffer from severe clinical manifestations of primary HHV6 infection. Asano et al. (1991) reported that the magnitude of virus replication in infants with roseola infantum was related to the duration and severity of clinical symptoms. Roseola infantum has been associated with febrile convulsions in some children and has been linked to some cases of SIDS (Berenberg et al., 1949). Encephalitis (Irving et a i, 1990, Ishiguro et al., 1990), hepatosplenomegaly (Irving et al., 1990) hepatitis (Asano et al., 1990a, Tajiri et al., 1990, Asano et al., 1991), bronchopneumonia (Asano et al., 1991) and intussusception (Asano et al., 1991b) have all been reported in young infants with roseola infantum and proven primary HHV6 infection. Haemophagocytic syndrome (a rare immunological disorder of children characterized by phagocytosis of red blood cells, leukocytes and platelets) has previously been associated with CMV, HSV and EBV infections (Olson and Olson, 1986). Huang et al. (1990) reported a case of fatal haemophagocytic syndrome in a child undergoing primary HHV6 infection.

It is possible that the HHV6 infection caused the severe clinical symptoms suffered by these children but results are difficult to interpret on single case studies. The HHV6 infection may have contributed to the severity of illness but since seroconversion for

186 HHV6 is common in children of the ages described, a causative role is difficult to prove.

Okano and coworkers (1989) undertook a study of HHV6 in children with Kawasaki disease (an acute febrile vasculitis of infancy). They reported increased anti-HHV6 IgG titres in children with Kawasaki disease when compared with controls. These results were not confirmed in studies by two other research groups (Enders et al ., 1990, Marchetteet al ., 1990).

4e. 2. sudden infant death syndrome

Many bacterial and viral infections have been associated with SIDS but there are probably many infectious and non-infectious causes of unexplained sudden death in a young infant.

Berenberg et al. (1949) first postulated that the causative agent of roseola infantum may be responsible for some cases of SIDS. In order to investigate a potential role for HHV6 in sudden infant death, tissues from cot death victims were screened for HHV6 DNA (and CMV DNA) by in situ hybridization. None of the lung, thymus, spleen and lymph node tissues screened were positive for HHV6 or CMV DNA by this method. Thus there was no evidence, from this study, for the association of these two viral infections with SIDS.

These results do not preclude association of HHV6 or CMV infections with some cases of sudden infant death. As tissues from 40 cases of SIDS were screened for HHV6 and CMV DNA, however, such an association is likely to be rare. It is possible that viral infection of the brain may cause some unexplained deaths in young infants but such tissue was not available for examination in this study.

4e. 3. chronic fatigue syndrome

Chronic fatigue syndrome has been defined as a clinical entity characterized by chronic debilitating fatigue lasting longer than 6 months (Komaroff, 1988). Patients with prolonged fatigue have historically been described as suffering from 'Icelandic disease', 'Royal Free disease', fibromyalgia, ME (benign or epidemic), PVFS or CFS. Holmes et al. (1988) attempted to unify the clinical diagnosis of CFS but the 'Holmes criteria' have not, as yet, been universally accepted for the classification of patients with chronic fatigue. The term PVFS has been used by clinicians to describe chronic fatigue when a definite 'acute phase'of illness (often flu-like symptoms) can be identified. Where brainstem signs are associated with prolonged symptoms of fatigue, patients are often classified as suffering from ME (Bell et al., 1988).

187 Epstein-Barr virus infection has been associated with some cases of PVFS (discussed in Buchwald and Komaroff, 1991). The isolation of HHV6 from patients with CFS (Ablashi et a l, 1988, Straus, 1988, Luka et al, 1990, Josephs et a l, 1991, Daugherty et a l, 1991) led to renewed research interest in the association of lymphotropic human herpesviruses with CFS.

In order to investigate any association of HHV6 infection with symptoms of prolonged fatigue, sera from two groups of patients were screened for HHV6-specific antibody. The first group of patients were described as having a glandular fever-like illness with recurrent fatigue that was not associated with a recent EBV infection (Read, et al., 1990). The PBNGF patients had a higher prevalence and titre of anti-HHV6 IgG, by indirect IF, than a group of age and sex matched controls. Antibody to CMV, HSV and EBV in these patients, however was not significantly different to that in controls.

There have been a number of reports of HHV6 infection associated with a mononucleosis-like illness (Krueger et a l, 1988a, Niederman et a l, 1988, Steeper et a l, 1990, Irving and Cunningham, 1990). Some of these patients progressed to a chronic disorder with fatigue and recurrent lymphadenopathy, similar to the patient described by Krueger et a l (1987). During the course of this study an adult (patient 20) was identified who had a persistent febrile illness and lymphadenopathy associated with serological reactivation of anti-HHV6. Anti-HHV6 specific IgM and IgG was demonstrated in serum samples from this patient taken during the course of his illness. Borisch Chappuis and colleagues (1989) reported a significant rise in anti-HHV6 IgG titre in a patient with pancreatic carcinoma who had symptoms of lymphadenopathy.

Eizuru et a l (1989) reported detection of HHV6-specific antigens by immunohistochemistry in 17 of 18 cervical lymph nodes from patients with histiocytic necrotising lymphadenitis. As 6 of 8 lymph nodes from 'control' patients with lymphadenitis or malignant lymphoma were also positive for HHV6 antigens this finding may not have been related to illness. The virus may persist in lymph node tissue of healthy individuals and reactivation may be associated with symptoms such as lymphadenopathy.

There are probably many viral causes of chronic fatigue. Raised antibody titres to a variety of viruses including EBV, HSV, CMV and measles have been described in some groups of CFS patients (discussed in Holmes et a l, 1987, Straus et al, 1985, Straus, 1988). In the group of PBNGF patients described in this study, anti-HHV6 IgG but not antibody to the other human herpesviruses was elevated indicating that these results were not due to non-specific polyclonal activation. The elevated anti-HHV6 in PBNGF patients may have been due to a secondary reactivation of the virus which was unrelated

188 to acute illness. It is possible, however that HHV6 had a direct role in the mononucleosis-like illness suffered by these patients.

The second group of patients with persistent fatigue who were screened for anti-HHV6 were enrolled into a CFS study group based on the clinical criteria described by Holmes et al. (1988). The prevalence and level of anti-HHV6 in this group of patients was not significantly different to that of age and sex matched controls. Thus, there was no evidence for an association of HHV6 infection with the symptoms of persistent fatigue suffered by this group of patients.

Differences in anti-HHV6 results when comparing the two groups of patients with chronic fatigue are probably accounted for by differences in clinical criteria for enrolment into study groups. Other research groups have reported no significant difference in prevalence or level of anti-HHV6 when comparing CFS (or ME) patients with controls (Wakefield et a l , 1988, Gold et al., 1990, Levy et al., 1990a, Smithet al., 1991, Marshallet al., 1991).

4e. 4. hepatitis

There have been a number of reports of HHV6 infection associated with hepatitis in children (Asano et al., 1990a, Tajiri et al., 1990, Asano et al., 1991) and adults (Kircheschet al., 1988, Niederman et al., 1988, Sobue et al., 1991). One child suffered fatal fulminant hepatitis during HHV6 primary infection. Although HHV6 has not been isolated from liver tissue, primary infection and reactivation of HHV6 in liver transplant recipients has been reported (Ward et al., 1989, Chou and Scott, 1990, Sutherland et al., 1991).

4e. 5. acquired immune deficiency syndrome

Many of the original HHV6 isolates were obtained from HIV infected individuals (Salahuddin et al., 1986, Downing et al., 1987, Tedder, et al, 1987, Lopez et al, 1988) and the virus has been shown to infect CD4 positive T-lymphocytes in vitro (Lusso et al., 1987, Ablashi et al, 1987). One group reported a higher prevalence and titre of HHV6 antibody in sera from HIV-1 infected than in sera from HIV-1 uninfected homosexual men (Krueger and Ablashi, 1987, Ablashi et al, 1988b). The authors postulated that HHV6 was a cofactor for the development of HIV related disease in these individuals.

The possible in vivo exacerbation of HIV infection by HHV6 was investigated during the course of this study. There was no significant difference in HHV6 antibody presence or level when sera from HIV-1 infected and uninfected homosexual men were

189 compared. There was also no difference in HHV6 antibody when anti-HIV-1 positive, p24 antigen positive and anti-HIV-1 positive, p24 antigen negative sera from homosexual men were compared. The HHV6 antibody in all groups of sera from homosexual men was not significantly different to that for male blood donors and thus there was no serological evidence for the potentiation of HIV infection by HHV6 in these individuals. A number of other independent serological studies was published which concluded that HHV6 infection was not a risk factor for HIV-1 seroconversion or progression to AIDS (Brown, et al ., 1988a, Holmberg et al. , 1989, Levy et al ., 1990a, Rodier et al ., 1990, Enders et al ., 1990, Ceccherini-Nelli et al ., 1990, Soriano et al., 1991). Those research groups that reported higher prevalence with or without increased titre of anti-HHV6 in HIV infected individuals (Ablashi et al ., 1988, Ablashi et a l, 1988a, Krueger et al. , 1988a, Huemer et al., 1989, Scott and Constantine, 1990) generally reported low prevalence and or titre of anti-HHV6 in the healthy adult population. Discrepancies in serological results reported by different research groups may be due to interpretation of a positive result by indirect IF.

Sera from HIV-1 infected and uninfected haemophiliacs were also screened for HHV6 antibody, this time by both indirect IF and competitive RIA. Again, there was no serological evidence for the potentiation of HIV infection by HHV6. Although the numbers in the haemophiliac groups were small the mean % inhibition by competitive RIA was significantly higher in anti-HIV-1 negative than in anti-HIV-1 positive sera from these individuals. A number of other research groups also reported lower prevalence and or titre of anti-HHV6 in sera from HIV antibody positive patients when compared with controls (Brown et al., 1988, Spira, et al., 1990, Ceccherini-Nelli et al., 1990). Holmberg et al. (1989) screened sera from HIV-1 antibody positive homosexual men for anti-HHV6. They reported that HHV6 antibody negative individuals were more likely to develop AIDS than HHV6 antibody positive individuals. Discordant anti- HHV6 results may be due to the clinical status of the HIV infected individuals studied. In many reports the p24 antigen or clinical status of the patients was either not known or not quoted.

There has been a number of in vitro studies to investigate potential interactions between HHV6 and HIV at the cellular level. The in vitro coinfection of CD4 positive lymphocytes by HIV-1 and HHV6 was first described by Lusso et al. (1989). Anti-CD4 monoclonals or soluble CD4 did not inhibit HHV6 infection of T-lymphocytes under conditions that did inhibit HIV infection (Lusso et al., 1989a). Thus it was concluded that infection of CD4 positive T-lymphocytes by HHV6 did not require the CD4 receptor. In vitro inhibition of HIV replication by HHV6 has been reported by a number of research groups (Pietroboni et al., 1988, Levy et al, 1990b, Agut et al., 1989, Carrigan et al., 1990). One group reported that this inhibition of HIV replication led to

190 prolonged survival of CD4 positive cells in culture (Levy et al, 1990b). Agut et al. (1989), however suggested that in vitro inhibition of HIV replication was due to destruction of lymphocytes by HHV6.

In contrast, there has been a report of productive dual infection of CD4 cells with HHV6 and HIV-1 which resulted in accelerated expression of HIV-1 genes and cell death (Lusso et al., 1989). Transactivation of the HIV promoter by HHV6 in chloramphenicol acetyltransferase assays has also been reported (Lusso et al, 1989, Ensoli et al, 1989, Horvat et al, 1989, Campbell et al, 1991, Horvat et al., 1991). In a recent study by one research group (Lusso et al, 1991, 1991a), HHV6 infection was reported to upregulate the expression of CD4 in a T-cell line and thus increase the number of cells available for infection with HIV. The CD8 positive cells in this culture were not productively infected with HIV unless simultaneously (or previously) infected with HHV6. Two other human herpesviruses (HSV-1 and CMV) did not induce the upregulation of CD4 expression in this tissue culture system. The authors suggested that if such a mechanism was active in vivo then the exacerbation of HIV infection by HHV6 may occur by an increase in the number of cells available for HIV infection and destruction. As the CD4 : CD8 ratio decreases with progression of HIV related disease it is difficult to see how this mechanism could be involved in the pathogenesis of HIV related disease in vivo.

Buchbinder et al. (1988) reported that peripheral blood samples from 52 of 63 AIDS patients were positive for HHV6 DNA by PCR. They did not report the prevalence of HHV6 DNA in healthy individuals, however and the significance of these results was not clear. Gopal and colleagues (1990) compared HHV6 and EBV DNA prevalence by PCR in peripheral blood and oropharyngeal samples from HIV infected and uninfected adults. The prevalence of HHV6 DNA was low in HIV infected individuals, particularly in those with symptomatic disease. The prevalence of EBV DNA in peripheral blood and oropharynx was significantly higher in HIV infected than in uninfected individuals.

Many of the published studies of HHV6 in HIV infected individuals are contradictory and make interpretation of these results difficult. When all the serological, molecular and cellular studies are taken into account the association of HHV6 with the pathogenesis of HIV related disease has not, as yet, been proven. Many viruses will transactivate the HIV promoter (including HSV, CMV and VZV, Rando et al., 1987) and could potentially interact with HIV in vitro.

As herpesviruses are ubiquitous and may reactivate as a result of immunosuppression, it is difficult to prove a role for these viruses in the exacerbation of HIV related disease. Infection with HIV alone may be sufficient for the depletion of CD4 positive cells, a

191 phenomenon which is characteristic of disease progression. A number of serological and isolation studies have linked herpesvirus (particularly CMV) infections with a more rapid progression to HIV related disease (Quinnan et al., 1984, Fiala et al., 1986, Webster et al., 1989). McKeating et al. (1990) also reported in vitro culture studies which indicated that the production of Fc receptors on the surface of CMV infected fibroblasts conferred susceptibility to infection with HIV.

Immunocompromised individuals may suffer from more severe primary or recurrent herpesvirus infections than healthy individuals. Herpesvirus infections are responsible for much of the morbidity and mortality of AIDS patients. In particular, CMV infection of AIDS patients may result in pneumonitis, gastrointestinal ulceration and hepatitis (reviewed by Griffiths, 1990). Patients with AIDS have an increased risk of development of B-cell lymphomas, some of which are associated with EBV (reviewed by Crawford and Edwards, 1990). A number of research groups have attempted to identify diseases associated with HHV6 infection in AIDS patients. Kaposi's sarcoma (a multifocal sarcoma of endothelial cell origin) is more common in AIDS patients than in other immunocompromised individuals and has been associated with CMV infection (discussed in Giraldo, et al., 1989). Jahan et al. (1989) analyzed cell lines derived from Kaposi's sarcoma for evidence of CMV and HHV6 DNA, none were positive. Using PCR, one research group identified HHV6 DNA in intraretinal lesions of 3 patients with AIDS associated retinitis (Qavi et al, 1989). As CMV and HIV were also detected in ophthalmic tissues from these patients it is difficult to determine any disease associated with HHV6 infection of the retina. Interactions between these three viruses may be important in the pathogenesis of retinal lesions.

4e. 6. disease and graft rejection in transplant recipients

A number of research groups have reported serological reactivation of HHV6 in transplant patients. Morris et al. (1989a) reported both primary infection and reactivation of HHV6 in renal transplant recipients. Three of the 4 transplant patients who suffered a non-specific febrile illness at the time of HHV6 infection also had serological evidence of primary CMV infection. The HHV6 infection may have been responsible for the febrile illness suffered by these patients and further studies are required to prove this association. Okuno et al. (1990) reported serological reactivation of HHV6 in 8 of 21 renal transplant recipients, in each case the HHV6 infection was associated with graft rejection.

Isolation of HHV6 from PBMC and kidney biopsy samples of renal transplant patients has been reported (Wrzos et al., 1990, Okuno et al., 1990, Asano et al., 1989). Kikuta et al. (1991) screened PBMC samples from renal transplant recipients for HHV6 DNA by

192 PCR. Eight of 9 patients samples were positive for HHV6 DNA compared with none of 5 samples from healthy adults. Ward et al. (1989) isolated HHV6 from PBMC and demonstrated seroconversion for anti-HHV6 IgM and IgG in a 21 year old patient suffering from neurological symptoms after liver transplantation. Primary infection and reactivation of HHV6 in liver transplant recipients has also been reported by two other research groups (Chou and Scott, 1990, Sutherland et al ., 1991) and one other group isolated HHV6 from a buffy coat sample from a liver transplant recipient (Luka et a l ., 1990).

Immunosuppressed patients often have concomitant HHV6 and CMV infections. Some research groups have used sera from these patients to assess any serological crossreaction between HHV6 and CMV. Morris et al. (1988) reported that although many transplant recipients had concomitant herpesvirus infections some patients with HHV6 infection remained anti-CMV negative. Absorption of sera with HHV6, CMV, EBV and VZV infected cells confirmed the lack of serological crossreaction between these viruses (Buchbinder et al., 1989). Chou and Scott (1990) also reported elevated anti-HHV6 in liver, heart and kidney transplant patients. They found that the highest anti-HHV6 titres occurred in patients with concomitant primary CMV infection but there was no correlation between CMV and HHV6 antibody titres in single serum samples. In serological studies of HHV6 and CMV in liver transplant recipients, however crossreactive antibodies were reported in some serum samples (Sutherland et al., 1991).

All these reports indicate that immunosuppressed renal and liver transplant recipients may reactivate HHV6, in some cases this may be associated with concomitant, CMV infection. There have also been reports of HHV6 primary infection after renal and liver transplantation when indirect IF was used for the detection of anti-HHV6. It is possible that some of these patients were anti-HHV6 positive prior to transplantation but that antibody levels were low and so undetectable by this technique. Infection of renal transplant recipients with HHV6 was associated with febrile illness and graft rejection in some cases. Clinical symptoms associated with HHV6 (but not CMV) infection of liver transplant recipients included fever, neurological disorders, hepatitis and lung dysfunction (Sutherland et al., 1991).

Recently, there has been a report of HHV6 associated pneumonitis in 2 adult bone marrow transplant recipients (Carrigan et al., 1991). In one case, HHV6 was isolated from bone marrow and peripheral blood samples and there was no evidence for any other concomitant infection which may have been associated with the pneumonitis (the patient was CMV seronegative). The patient suffered recurrence of his carcinoma and poor bone marrow function and subsequently died. Immunohistochemical staining of

193 lung tissue from this patient taken at autopsy revealed widespread HHV6 infection of the lung.

In the second patient with pneumonitis, HHV6 was isolated from broncheolar lavage and sputum and HHV6-specific antigens were detected at a time when no other infectious cause of illness was found. Later CMV was detected in peripheral blood samples and the patient was treated with forscarnet and subsequently gancyclovir. During treatment, HHV6 infected cells were identified in bone marrow samples but after treatment the pnemonitis resolved. The patient, however remained leucopenic and died from disseminated adenovirus infection. It is tempting to speculate that HHV6 infection was the primary cause of pneumonitis in this patient and that the HHV6 infection was responsible for the subsequent reactivation of CMV. In vitro studies have suggested that the susceptibility of HHV6 to anti-viral drugs is similar to that of CMV ( Russler et al., 1989, Agut et al., 1989, Burns and Sandford, 1990, see 4h.). Thus the treatment of this patient with forscarnet and gancyclovir may have inhibited HHV6 and CMV infections.

Pneumonitis caused by HHV6 infection is probably rare in bone marrow transplant recipients. Yoshikawa etal. (1991) studied 25 pediatric bone marrow transplant patients for evidence of HHV6 infection. Virus was isolated from PBMC and or bone marrow of 10 of these patients and 2 other patients had a significant increase in HHV6 antibody titre. Four patients had mild symptoms of skin rash and or fever at the time of HHV6 infection. Carrigan et al. (1991) reported no isolation of HHV6 from bone marrow and lavage samples of a group of BMT recipients followed prospectively. The authors did report isolation of HHV6 from PBMC from a bone marrow transplant patient during an episode of meningoencephalitis. This may be a rare complication of HHV6 infection in immunosuppressed adults and has been reported in a child with roseola infantum (Ishiguro et al., 1990). Asano et al. (1991a) reported isolation of HHV6 from PBMC of 3 pediatric bone marrow transplant recipients, 2 of these patients suffered rash and fever at the time of isolation.

4e. 7. benign lymphoproliferative disorders

Polyclonal lymphoproliferation has been reported in associated with HHV6 infection (Krueger et al., 1988) and the potential association of HHV6 with benign lymphoproliferative disorders has been studied by a number of research groups. Biberfeld and coworkers (Biberfeld et al., 1988, Ablashi et al., 1988) reported higher prevalence and titre of anti-HHV6 IgG in patients with acute sarcoidosis (characterized by the formation of epithelial granulomas in various organs) than in healthy individuals. The authors also reported in situ localization of HHV6 DNA in 1 of 5 lymph node

194 tissues from patients with this disease. As the prevalence of HHV6 in normal lymph node tissue was not reported by these authors the significance of these results remains unclear. Elevated anti-HHV6 titres were not found in sera from patients with rheumatoid arthritis (Biberfeld, e ta l, 1988, Baboonian et al., 1990).

The isolation of HHV6 from saliva samples of healthy individuals led to much research interest in a potential association of HHV6 with SS. One group reported higher prevalence and titre of anti-HHV6 IgG in patients with SS when compared with controls (Biberfeld et al, 1988) but these results were not confirmed in a second report (Baboonian etal., 1990).

In the study undertaken in this thesis, in situ localization of HHV6 proteins in parotid salivary glands from SS patients was undertaken. All of the 10 tissue samples screened by immunohistochemistry were positive for HHV6 proteins. Interestingly, the mononuclear cells infiltrating the glands from SS patients were all positive for HHV6- specific proteins but were not stained with the control (rubella and CMV) MAbs. Krueger and colleagues (1990) reported similar results for lip salivary glands from SS patients. Thus it is possible that HHV6 has a role in the pathogenesis of this disease.

Fox et al. (1986, 1987) reported detection of EBV early antigen in ductal and or acinar cells of 8 of 14 salivary glands from SS patients but not in salivary gland from healthy individuals or patients with salivary gland tumours. Epstein-Barr virus-specific DNA was also detected, by Southern and slot blotting, in some tissue extracts and parotid saliva samples from SS patients. The authors postulated that SS results from an abnormal immune response to a ubiquitous virus. They suggested that expression of viral antigens on the surface of salivary gland cells may lead to immune T-cell destruction of these cells. This postulate could apply to EBV or HHV6 infection. More work needs to be done to determine any interaction between these (and other viruses) which may be important in the pathogenesis of SS.

4e. 8. tumours derived from lymphoid and nervous tissue

Association of the HHV6 genome and subgenomic DNA fragments with neoplastic transformation has been shown by in vitro transformation of cell lines and induction of rapidly growing tumours in nude mice (Razzaque, 1990). These results confirmed that HHV6, like the other human herpesviruses, has oncogenic potential and may have a role in the pathogenesis of natural occurring human neoplasias.

Serological studies have associated HHV6 infection with the pathogenesis of lymphoid malignancies. Ablashi et al. (1988) reported higher prevalence and titre of anti-HHV6 IgG in patients with Hodgkin's disease, lymphocytic leukemia and Burkitt's lymphoma.

195 Clark et al. (1990) also undertook a serological study of HHV6 in leukaemias and lymphomas and reported a higher prevalence and titre of anti-HHV6 IgG in patients with acute myeloid leukaemia, Hodgkin's disease and low-grade non-Hodgkin's lymphoma when compared with age and sex matched controls. Treatment induced immunosuppression did not influence the presence or titre of HHV6 antibody. Another study (Torelli et al., 1991) also indicated that anti-HHV6 titres were higher in patients with Hodgkin's lymphoma than in controls.

Two groups reported the detection of HHV6 DNA, by Southern blotting, in a small proportion of B- and T-cell lymphomas (Josephs et al., 1988a, Jarrett et al., 1988). In two of the patients with HHV6 DNA positive B-cell lymphomas the malignancy occurred in the context of SS. Using PCR amplification, Buchbinder et al. (1988) reported that 20 of 23 tumours contained HHV6-specific DNA and in one case (Burkitt's lymphoma) it was possible to localize the viral DNA by in situ hybridization. Luka and colleagues (1991) detected HHV6 DNA by PCR and in situ hybridization in T-cells from children with the rare T-cell acute lymphoblastic leukaemia. As no late HHV6 proteins were detectable by staining with MAbs the authors concluded that HHV6 was latent in these lymphoblasts and may be involved in the aetiology of this malignancy. Torelli et al. (1991) reported that 3 of 25 tissue samples from Hodgkin's lymphoma were positive for HHV6 DNA by PCR.

In the study reported in this thesis, approximately half of the gut lymphoma tissues screened by nested PCR were positive for HHV6 DNA. The ubiquitous nature of HHV6, the extreme sensitivity of PCR and the demonstration of virus in histologically normal salivary gland and large bowel tissues led to the conclusion that HHV6 is rarely, if ever, associated with the pathogenesis of these malignancies. Virus-specific DNA was also detected in nucleic acid extracts from 1 neuroblastoma and 2 medullablastomas but again the low prevalence of HHV6 in these tissues would indicate that an association of HHV6 with the pathogenesis of these malignancies is rare. In situ localization of HHV6 DNA and or protein within these malignant tumours would help to confirm a role for the virus in neoplastic disease.

4e. 9. inflammatory bowel disease

As part of a study of viral infections associated with IBD, large bowel tissue extracts from patients with UC, CD and patients with a non-inflammatory disease were screened for HHV6 DNA by nested PCR. The prevalence of HHV6 DNA was high in patients and controls indicating that HHV6 infection of the large intestine is common. This was the first report of HHV6 infection of intestinal tissue. Levy and coworkers (1990) had

196 previously reported productive infection of cell lines derived from large bowel tissue by the HHV6SF isolate.

In order to interpret the high prevalence of HHV6 in large bowel tissues, nested PCR for the detection of HSV 1-, VZV-, CMV- and EBV-specific DNA was developed. None of the tissues from patients or controls contained HSV-1 or VZV DNA but the prevalence of CMV and EBV DNA was significantly higher in tissues from CD and UC patients than in control tissues. The detection of herpesvirus DNA in large bowel tissues was unrelated to corticosteroid treatment.

The association of CMV infection with some cases of UC has been reported previously (Farmer et al ., 1973, Morton, 1985, Roberts et al., 1989). In the study reported in this thesis the most interesting finding was that multiple infection of large bowel tissue with HHV6 and or CMV and or EBV was significantly higher in UC patients than in CD patients and controls. Interactions between these three herpesviruses may be important in the pathogenesis of UC and requires further study. In particular, the in situ localization of HHV6 in normal and inflammed bowel tissue may confirm association of viral replication with diseased areas of tissue.

The high levels of CMV and EBV DNA in tissues from both CD and UC patients, but not controls, may indicate association of these viruses with the inflammatory response which is characteristic of both CD and UC.

It has been suggested that CD and UC may be caused by an autoimmune mechanism. There is much evidence of immunological abnormalities in IBD which may be important in the perpetuation of the disease state (Jewell and Snook, 1990). Roche and Huang (1977) first suggested that immune-mediated cytolysis of virus infected cells may be the pathogenic mechanism involved in inflammatory bowel disease. The reduction of immune cytolysis as a result of the immunosuppressive nature of steroids may account for the reduction in the severity of symptoms seen with this treatment.

4f. Interaction of HHV6 with the other human herpesviruses

Studies of children with roseola infantum have confirmed the association of this disease with primary HHV6 infection. In most cases of roseola infantum there is no evidence of any other active infection which may be responsible for the clinical symptoms. There has been a report of concomitant HHV6 and measles primary infection of 2 children (Suga et al., 1990). These children suffered from symptoms which were atypical of either roseola infantum or measles. Irving et al. (1990a) identified 2 children with primary HHV6 infection who had concomitant primary EBV or CMV infections. The clinical symptoms suffered by these children may have been more severe because of the

197 dual infection. A 3 year old child with primary EBV infection associated with serological reactivation of HHV6 has been described in this thesis (patient 19).

Anti- HHV6 serological results for adult patients are difficult to interpret. Active HHV6 infection in adults without evidence of a concomitant infection with one of the other human herpesviruses has been reported (patient 20, Morris et al ., 1988, Niederman et al ., 1988, Steeper et al ., 1990, Irving and Cunningham, 1990, Irving et al ., 1990, Sutherland et al, 1991). Raised levels of anti-HHV6 IgG with or without the presence of specific IgM antibody have also been demonstrated in patients with primary or reactivated CMV, EBV or VZV infections.

Although many research groups have addressed the potential problem of serological crossreaction between HHV6 and the other human herpesviruses there has been only 1 study which has demonstrated this phenomenon (Sutherland et al ., 1991). These authors absorbed sera from liver transplant recipients with HHV6 and CMV antigens and confirmed that crossreactive antibodies were produced in some of these patients.

The balance of serological evidence would indicate that multiple HHV6 and or CMV and or EBV infections occur quite commonly. Serological reactivation of anti-HHV6 may occur in response to a primary CMV infection and HHV6 and CMV may reactivate concurrently. In the latter case infection with one virus may reactivate the other or the two viruses may reactivate in response to another stimulus. Similar interactions between HHV6 and EBV may occur under some circumstances.

Chronic fatigue syndrome has been associated with EBV infection but isolation of this virus from patients with prolonged fatigue has not been reported. Isolates of HHV6 have been obtained from PBMC of patients classified as suffering from CFS (Ablashi et al ., 1988, Straus, 1988, Josephs et al ., 1991, Daugherty et al ., 1991). Two groups (Ablashi et al ., 1991, Daugherty et al ., 1991) reported serological evidence of concomitant HHV6 and EBV infections in some CFS patients and suggested that interactions between these viruses may be important in the pathogenesis of disease.

Tissue localization of HHV6 has revealed that the virus persists in salivary gland and large bowel tissues, both of which may be coinfected with CMV and or EBV. Sjogrens syndrome has most frequently been associated with EBV infection but EBV-specific DNA and protein have not been detected in salivary gland tissue from all these patients (Fox et al ., 1986, 1987). The localization of HHV6-specific antigens in columnar epithelium, mucous, serous and mononuclear cells in salivary glands from SS patients may indicate an association of HHV6 with the pathogenesis of this disease. Coinfection of salivary gland tissue with HHV6 and EBV may be important in the disease process.

198 The association of multiple herpesvirus infections with UC requires further investigation. Again, HHV6, CMV and EBV have all been associated with this disease. Treatment of UC patients with corticosteroids reduces the severity of clinical symptoms but often gives only a temporary respite. The long term effects of steroid treatment on in vivo herpesvirus infections requires further study.

4g. Natural history and pathogenesis of HHV6 infection

Serological reactivation of anti-HHV6 during pregnancy has been demonstrated. The number of women shedding HHV6 in the throat and the amount of virus shed was found to increase with advancing gestation. The high levels of HHV6 in oropharyngeal samples were maintained in the early postnatal period.

Most children are infected with HHV6 during the first year of life. Clinical symptoms associated with primary HHV6 infection may either be mild, inapparent or manifest as the febrile illness known as roseola infantum. There is no evidence for congenital acquisition of HHV6 nor for infection at birth due to contact with virus shed in the birth canal. Roseola infantum occurs as sporadic cases and not in epidemic form and so horizontal transmission of the virus probably occurs from infected adults to children. Breast milk is not a common route of transmission of the virus and so transmission by saliva seems likely. As the presence and level of HHV6 in TW is higher in women postpartum than in non-pregnant adults it is resonable to assume that many infants acquire HHV6 from the saliva of their mothers.

The early acquisition of HHV6 would suggest that transmission occurs, at least in some cases, in the presence of significant levels of anti-HHV6 maternal antibody. Humoral immunity may modulate clinical symptoms but not prevent primary infection. Localized infection of salivary gland tissue probably leads to vascular spread of the virus within PBMC or cell free in plasma. After the initial viraemic phase the amount of HHV6 in peripheral blood drops considerably unless reactivation occurs but the virus may persist in lymph node and other lymphoid tissue. Monocytes or macrophage may be a site of latency of HHV6.

Infection of tissues of the large bowel probably occurs during the viraemic phase of infection. It is possible, but seems unlikely, that vims ingested during the initial transmission event passes through the stomach and small intestine and infects colonic tissue.

The in vivo tissue tropism of HHV6 seems to be wide and infection of major and minor salivary glands, kidney, lung, lymph node and colonic tissues have been reported. In some cases HHV6 infection has been associated with hepatosplenomagaly, hepatitis and

199 encephalitis which may indicate infection of liver and neurological tissue. We do not, as yet, know the full spectrum of tissues which may be infected with HHV6 in healthy individuals and patients with disease.

Reactivation of HHV6 seems to occur frequently in adults either in response to other viral infections or in patients without any evidence of a concomitant infection. In immunocompetent individuals HHV6 infection may be associated with a mononucleosis-like illness with or without symptoms of persistent fatigue. Localized reactivation and replication of HHV6, perhaps in response to hormonal stimuli, may account for the increased shedding of the virus in saliva of women during pregnancy and the early postnatal period. There may be an increase in the amount of virus in peripheral blood in adults suffering from HHV6 infection, accounting for the isolation of HHV6 from many of these individuals. Reactivation of HHV6 may occur in some individuals who are immunosuppressed (either iatrogenically or due to other infections).

4h. Prevention and treatment of HHV6 infection

Studies have been undertaken to determine the in vitro susceptibility of HHV6 to anti viral drugs. The viral thymidine kinase (TK) encoded by HSV 1, HSV 2, VZV and EBV efficiently activates (phosphorylates) acyclovir and bromovinyl-deoxyuridine (BVdU), thus they are sensitive to low doses of these anti-viral drugs. Cytomegalovirus does not encode a TK gene and is relatively resistant to acyclovir and BVdU.

Most studies have shown that HHV6 replication is largely unaffected by acyclovir and BVdU unless high concentrations of these anti-viral drugs are used (Streicher et al ., 1988, Russler et al., 1989, Agut, et al, 1989a, Di Luca et al., 1990, Bums and Sandford, 1990, Agut et al., 1991). Differences in the isolates of HHV6 or the conditions used for the in vitro studies may explain why another research group (Kikuta et a l, 1989a) reported inhibition of HHV6 replication with lower doses of acyclovir. These results either indicate that HHV6 does not encode a viral TK or that the HHV6 DNA polymerase is less susceptible to the phosphorylated form of these anti-viral drugs than HSV, VZV and EBV. Agut et al. (1991) reported that the in vitro susceptibility of different HHV6 isolates to antiviral drugs was similar.

In vitro studies have indicated that HHV6 shuts off host DNA synthesis early in infection (Di Luca et al., 1990), whereas CMV infection transiently activates host macromolecular synthesis. It has been suggested that this activation upregulates cellular TK activity enabling efficient uptake of thymidine by CMV (Griffiths, 1990). If HHV6 does not encode a TK gene then other methods for uptake of thymidine must be active in infected cells.

200 Gancyclovir, phosphonoformic acid (forscamet) and phosphonoacetic acid do not require a fuctional viral TK for phosphorylation and are active against CMV. One group reported that gancyclovir was toxic to uninfected cells and had little effect on the replication of HHV6 (Streicher et al., 1988) but other studies have confirmed the inhibition of HHV6 replication in vitro by this anti-viral drug (Russler et al., 1989, Agut et al., 1989, Bums and Sandford, 1990). A number of research groups have also shown the selective inhibition of HHV6 replication in vitro by phosphonoformic acid and phosphonoacetic acid (Streicher et al., 1988, Agut, et al., 1989, Di Luca et al., 1990, Bums and Sandford, 1990). In 1 study 9-[4-hydroxy-2-(hydroxymethyl)butyl] guanine was found to be a more potent inhibitor of HHV6 replication than gancyclovir (Akesson-Johansson et al., 1990). The majority of in vitro studies have indicated that the pattern of susceptibility of HHV6 to anti-viral drugs is similar to CMV.

There have been only two reported cases of treatment of a patient with confirmed HHV6 infection. A bone marrow transplant recipient with pneumonitis associated with concomitant HHV6 and CMV infections was treated with forscamet and gancyclovir (Canigan et al., 1991) with a subsequent remission of respiratory symptoms. A second immunocompetent patient had concomitant Legionella and HHV6 infections associated with respiratory, hepatic and CNS disease (Russler et al., 1991). Antibiotics alone were unsuccessful in the treatment of this patient but high dose corticosteroids and antibiotics cleared the Legionella infection. As in vitro studies had shown that replication of HHV6 was inhibited by corticosteroids, the authors concluded that the HHV6 infection may have exacerbated this patient's symptoms.

In view of the reported effect of corticosteroids on the replication of HHV6 (Russler et al, 1991) it is interesting that corticosteroids are frequently used to reduce the severity of symptoms associated with IBD, although such treatment may give only temporary respite. The incidence of colorectal cancer is higher in patients with a history of UC than in the general population (discussed in Gyde, 1990). If the association of herpesviruses with UC is confirmed it may be important to study more extensively the effects of corticosteroids on these viruses in vivo.

Induction of interferon and other cytokines has been reported in HHV6 infected mononuclear cells (Kikuta et al., 1990a, Flamand et al., 1991). This may be important in the host response to HHV6 infection and could potentially be used to treat such infections.

If an association of HHV6 infection with serious disease is confirmed it may become important to develop methods for prevention of HHV6 primary infection or reactivation. The methodology used for prevention and treatment of HSV, VZV, CMV

201 and EBV infections may be useful for the modulation of HHV6 infection. Live attenuated HSV, CMV and VZV vaccines have been prepared but are not, as yet, used routinely. Further studies are required to establish the frequency of latency and reactivation of these attenuated viruses. The oncogenic potential of herpesvirus DNAs has led to development of sub-unit HSV and EBV vaccines which do not contain viral DNA but results from studies utilizing these vaccines are not as yet available. Hyperimmune plasma and globulin have been given to immunosuppressed individuals in order to reduce clinical severity of herpesvirus infections.

4i. HHV6 molecular studies

While HHV6 epidemiological studies were underway, detailed analysis of the HHV6 genome was attempted by other research groups. These studies were used to define the evolutionary relationship between HHV6 and the other human herpesviruses. Most of the molecular studies of HHV6 were carried out using the GS (Salahuddin et al., 1986), U1102 (Downing et a l, 1987) and Z29 (Lopez et al., 1988) isolates of HHV6.

Initial studies by Salahuddin and colleagues (1986) showed a lack of homology between HHV6 (GS) DNA and that of the other human herpesviruses. Later studies showed that a 5.5 Kbp portion of the HHV6 (U1102) genome would cross hybridize with human CMV under high stringency conditions (Efstathiou et al., 1988). When a 21.8 Kbp portion of the HHV6 genome was sequenced, open reading frames homologous to a set of genes conserved in all other sequenced herpesviruses were identified (Lawrence et al., 1990). There was a high degree of sequence homology between HHV6 and CMV across this region. When the genetic structure of the human herpesvirus genomes were compared a rearrangement of the major capsid protein (MCP) gene was noted in both CMV and HHV6 when compared with the other human herpesviruses. Sequence analysis confirmed that the HHV6 MCP showed closer homology to CMV than to any of the other human herpesviruses (Lawrence et al ., 1990, Littler et al., 1990). Teo and colleagues (1991) sequenced the region of the HHV6 (U1102) genome around the DNA polymerase and confirmed that the organization of this region was also identical to CMV and different from the other human herpesviruses.

Restriction enzyme mapping of the complete HHV6 genome gave precise information on the size of the ds DNA molecule (161.5 Kbp) and was used to further deliniate the genetic structure of the virus (Lopez and Honess, 1990, Thomson et al., 1990, Martin et al., 1991). The mononucleotide composition of the HHV6 genome was 43 - 44 % G+C as estimated by buoyant density in caesium chloride. The structure of the DNA was found to resemble superficially that of channel catfish virus and equine

202 cytomegalovirus with a unique sequence (141 Kbp) flanked by a direct repeat sequence (10 Kbp) at each terminus of the DNA molecule (class A, as defined by Roizmann, 1990).

Multiple copies of the GGGTTA sequence in the HHV6 genome were reported and shown to be responsible for hybridization of HHV6 DNA to regions of the Marek's disease virus genome (Kishi et al., 1988). Martin and coworkers (1991) confirmed that this sequence was reiterated in the HHV6 unit-length genome at the junction of the repeat and unique sequences. The authors suggested that this sequence may be important in the cleavage of unit length HHV6 DNA from concatemeric molecules. They also noted that the GGGTTA sequence is highly conserved in telomeres of all vertebrate chromosomes studied (Meyne et al., 1989). It is possible that this repeat sequence was acquired during integration of HHV6 into the host cell genome (Dr R. Honess, personal communication) although there is not, as yet, any further evidence for this. A gene in HHV6 encoding a polypeptide homologous with the adeno-associated virus type 2 rep gene has also been reported (Thomson et al., 1991), the significance of this is not known.

4j. Classification of HHV6

Historically, herpesviruses have been classified into sub-families according to their biological characteristics. In most cases herpesviruses classified in the same sub-family on biological criteria are closely related at the molecular level. Although HHV6 is a lymphotropic human herpesvirus, analysis of the genetic organization and DNA sequence of the viral genome has indicated that HHV6 is more closely related to CMV than to any of the other human herpesviruses. Based on molecular studies it has been suggested that HHV6 should be classified with human CMV in the Betaherpesvirinae sub-family (Lawrence et al., 1990, Lopez and Honess, 1990) and not in the Gammaherpesvirinae sub-family with the other lymphotropic herpesviruses.

The results of the epidemiological studies reported in this thesis and other published reports have indicated that HHV6 and CMV also share many biological characteristics. The only cells fully permissive for CMV in vitro are human fibroblasts but in vivo the virus preferentially infects visceral epithelial cells. Cytomegalovirus commonly infects salivary gland, lung, kidney, liver, spleen and haematopoietic cells. Preliminary studies indicate that HHV6 will infect T-lymphocytes and human fibroblasts in vitro (4a.), exhibits a wide in vivo tissue tropism and infects many of the cells and tissues that may be productively infected with CMV. Mononuclear cells have been suggested as a potential site of CMV and HHV6 latency. Cytomegalovirus and HHV6-specific DNA may be detected in mononuclear cells of some healthy individuals by the sensitive PCR

203 technique. Neither virus is easily isolated from the peripheral blood of healthy individuals but during HHV6 or CMV primary infection or reactivation the ensuing viraemia allows viral isolation from peripheral blood. Three research groups (Russler et al ., 1989, Agut et al., 1989, Burns and Sandford, 1990) reported that the in vitro susceptibility of HHV6 to anti-viral drugs was similar to CMV and suggested that the two viruses may have other characteristics in common.

In contrast, the in vivo tissue tropism of the other human herpesviruses is fairly limited. Only lymphoid and nasopharyngeal epithelial cells have been identified as permissive for EBV infection in vivo, HSV predominantly infects squamous epithelium of skin and mucous membranes and VZV infects epidermal cells and the peripheral nervous system. All three viruses are susceptible to the nucleoside analogue acyclovir whereas in vitro studies have indicated that HHV6, like CMV, is relatively resistant to this drug.

At least some isolates of HHV6 may be propagated in PBMC from chimpanzees (Levy et al ., 1990, Lusso et al., 1990). Higashi et al. (1989) reported detection of anti-HHV6 in sera from chimpanzee and other monkey species. These results suggest that HHV6, or a closely related virus, may infect monkey species. Although a characteristic of most members of the Betaherpesvirinae is a restricted host range CMV related viruses have been identified in other animal species.

The frequent identification of HHV6, CMV and EBV in the same tissue sample (salivary gland and large bowel) may indicate a close relationship between these three viruses. Concomitant HHV6 and CMV and or EBV infections have been reported and further studies are required to determine any in vivo interactions between these viruses. Serological reactivation of anti-HHV6 in response to primary CMV and EBV infections has been described in this thesis and reported in the literature.

Antigenic diversity amongst different HHV6 isolates has been reported (Wyatt et al., 1990, Schirmeret al., 1990). Recently a seventh human herpesvirus (HHV7) has been described (Frenkel et al., 1990) which although clearly related to HHV6 seems to show only limited genomic and antigenic homology with the Z29 and U1102 isolates. Preliminary studies have indicated that this virus also shows tropism for CD4 positive T-lymphocytes but is immunologically distinct from HHV6 (Wyatt et al., 1991). More extensive studies will need to be undertaken in order to confirm the classification of the most diverse HHV6 isolates and the relationship to HHV7 and the other human herpesviruses.

204 4k. Conclusions

Human herpesvirus 6 is ubiquitous and most individuals are infected with the virus during the first year of life. As many women have high levels of HHV6 in saliva postpartum horizontal transmission of the virus from the mother to newborn offspring probably occurs frequently. Clinical symptoms associated with HHV6 primary infection are usually mild although there may rarely be serious complications in a young infant.

Antibody reactivation occurs commonly in adults, particularly in those who are immunosuppressed. In immunocompetent adults, a mononucleosis-like illness with or without symptoms of chronic fatigue may accompany HHV6 infection. Transplant recipients may suffer from more severe disease associated with HHV6 infection. Primary infection or reactivation of HHV6 in renal transplant recipients has also been associated with graft rejection.

Reactivation of HHV6 may occur in response to a concomitant CMV or EBV infection. Interactions between these three viruses may be important in the pathogenesis of diseases such as UC and SS and require further study.

At the molecular level, HHV6 is closely related to CMV but further information is required, particularly in light of the identification of a seventh human herpesvirus, before formal classification of these viruses is undertaken.

205 Appendix A: Media and buffer recipies

All media and buffers were made up in sterile distilled water unless otherwise stated.

Media

Media constituents were either bought as sterile stock solutions (RPMI, NaHC03, HEPES, glutamine, hypoxanthine, aminopterin, thymidine, FCS) or made as a stock solution and sterilized by filtration through Millipore 0.2 p. membranes (penicillin, streptomycin, fungizone, 2-mercaptoethanol). Media and supplements were obtained from Flow Laboratories Ltd. unless otherwise stated.

Complete RPMI Incomplete MAb medium (IM) 2.0 X 103 M L-glutamine 1 X RPMI 1640 (from 10X stock) 100 units / ml penicillin 2.4 X 10-2 M NaHC03 (Glaxo Laboratories Ltd.) 5.0 X 10-3 M HEPES buffer 100 pg / ml streptomycin (Evans Medical Ltd.) HT medium 2.5 pg / ml fungizone 1.0 X 10-4 M hypoxanthine (Squibb Ltd.) (Sigma Chemical Co.) 10 % (v/v) FCS 1.6 X 10'5 M thymidine Make up in IX RPMI (Sigma Chemical Co.) (Dutch modification). Make up in CM.

Complete MAb medium (CM) HAT medium 1 X RPMI 1640 (from 10X stock) 1.0 X 10-4 M hypoxanthine 5.0 X 10-3 M HEPES buffer 4.0 X 10‘7 M aminopterin 2.0 X 10-3 M L-glutamine (Sigma Chemical Co.) 5.0 X 10'5 M 2-mercaptoethanol 1.6 X 10'5 M thymidine (Koch-Light Ltd.) Make up in CM. 2.4 X 10*2 M NaHC03 100 units / ml penicillin Freezing medium 100 pg / ml streptomycin 10 % (v/v) DMSO 2.5 pg / ml fungizone (Analar, BDH Ltd.) 20 % (v/v) FCS 30 % (v/v) FCS Make up in IX RPMI (Dutch modification).

206 Buffers for serological studies

Reagents were Analar grade obtained from BDH Ltd. unless otherwise stated. Formulae are given for the simple compounds, the more complex buffer constituents are listed under their common names.

Phosphate buffered saline Tris-saline buffer (10X TSB) (PBS) 2.0 X 10-1 M Tris 1.5 X 10-1 M NaCl 1.5 X 10'1 M NaN3 1.5 X 10-3 M KH2P0 4 1.5 M NaCl 6.3 X 10-3 M Na2HP04.12H20 pH 7.6 2.6 X 10-3 M KC1 pH 7.2 Veronal buffer (5X VB) Supplied in tablet form 1.5 X 10-1 M NaCl (Difco Laboratories Ltd.). 2.7 X 10 2 M sodium barbital 4.0 X 10-3 M MgCl2.6H20 Phosphate buffer (200 mM) 8.0 X 10-4 M CaCl2.2H20 A = 2.0 X 101 M KH2P04 pH 7.2-7.4 B = 2.0 X 101 M Na2HP04 Supplied in tablet form Take 200ml B and adjust (Difco Laboratories Ltd.). to pH 8 with A. Dry acetone Tris-azide buffer (10X TAB) Shake acetone (99.5 %) with 2.0 X 10-1 M Tris Na2S04 (Sigma Chemical Co.) 1.5 X 101 M NaN3 until precipitate forms. pH 7.6 Filter through Whatman 1mm paper. Tris-BSA quench IX TAB (from 10X stock) Aminoalkylsilane solution 0.2 % (w/v) BSA 2 % (v/v) 3-aminopropyl- (Wilfrid Smith Ltd.) triethoxysilane (98 %) pH 7.6 (Aldrich Chemical Co.) Make up in fresh dry acetone. Antigen buffer just prior to use. IX TAB (from 10X stock) 0.5 % (w/v) BSA Buffer for diluting radiolabel pH 7.6 IX TAB (from 10X stock) 2 % (w/v) BSA pH 7.6

207 Mayer's haemalum DAB substrate mix (haematoxylin) 0.06 % (w/v) DAB 1.1 X 10-1 M A1K(S04)2.12H20 (Sigma Chemical Co.) 0.1 % (w/v) haematoxylin 5.0 X 102M Tris (Sigma Chemical Co.) 0.03 % (w/v) H20 2 5.1 X 10^ M sodium iodate (from 30 % stock) 4.8 X lO-3 M citric acid pH 7.6 monohydrate Make just prior to use. 3.0 X 1 0 1 M chloral hydrate

Buffers for molecular studies:

All reagents used for in situ hybridization and PCR were analytical grade from BDH unless otherwise stated.

Carnoy's fixative 20X SSC 60 % (v/v) ethanol (99.5%) 3.0 M NaCl (Hayman Ltd.) 3.0 X 1 0 1 M sodium citrate 30 % (v/v) chloroform (99.5%) pH 7.0 10 % (v/v) acetic acid (99.5%) Sterilize by autoclaving.

Sonicated salmon sperm DNA In situ hybridization mix 10 mg / ml salmon sperm DNA 50 % (v/v) deionised HCONH2 (Sigma Chemical Co.) 10 % (w/v) dextran sulphate Mix overnight at 4 °C. 2X SSC (from 20X stock) Syringe through 18G needle, 1.0 X 10-4 M EDTA boil for 10 minutes and 5.0 X 10-4 M Tris cool on ice. Syringe, boil, 100 jig / ml sonicated salmon and cool again. sperm DNA Store aliquoted at -20 °C. pH 7.5 Store aliquoted at -20 °C.

Deionised formamide TE buffer Stir HCONH2 (Gibco-BRL Ltd.) 1.0 X 10-2 M Tris with Biorad AG 501-X8 mixed 1.0 X 103M EDTA bed resin for 1 hour and pH 8 filter through Whatman 1mm Sterilize by autoclaving. paper. Store aliquoted at -20 °C.

208 In situ buffer 1 NBT / BCIP substrate mix 1.0 X KHM Tris 0.03 % (w/v) NBT 1.0 X 101 M NaCl (Sigma Chemical Co.) 2.0 X 10-3 M MgCl2.6H20 0.02 % (w/v) BCIP 0.05 % (v/v) Triton-X-100 (Sigma Chemical Co.) pH 7.5 1 X 10'3 M levamisole Make fresh and filter (from stock) through 0.45 p Dissolve NBT in 70% (v/v) Millipore membrane. DMF (Sigma Chemical Co.) and BCIP in DMF prior to addition In situ buffer 2 to substrate mix. Make 3 % (w/v) BSA in buffer 1 substrate up just prior to Make fresh and filter use in NBT / BCIP buffer. through 0.45 p Millipore membrane. Hexazotized new fuchsin 5 % (w/v) new fuchsin NBT / BCIP buffer (Sigma Chemical Co.) 1.0 X 1 0 1 M Tris 2.2 M concentrated HC1 1.0 X 10-!M NaCl Warm gently, cool and filter 5.0 X 10-2 M MgCl2.6H20 through Whatman 1mm paper. pH 9.5 Store at 4 °C in the dark. Make fresh and filter Just prior to use mix 50:50 through 0.45 p Millipore with fresh 6.0 X 102 M membrane. NaN02.

Levamisole solution New fuchsin substrate mix 1.0 X 1 0 1 M levamisole 0.02 % (w/v) napthol AS-TR (Sigma Chemical Co.) phosphate Make up in NBT/BCIP buffer. 0.03 % (w/v) hexazotized Filter through 0.45 p new fuchsin Millipore membrane. Dissolve napthol AS-TR Store aliquoted at -20 °C. phosphate in DMF prior to addition to substrate mix. New fuchsin buffer Make substrate up in new 2 X 1 0 1 M Tris pH 9.0 fuchsin buffer just prior to Make fresh and filter use and filter through through 0.45 p Millipore Whatman 1 mm paper. membrane.

209 PCR extraction buffer (EB) TAE (50X) 1.0 X 10-2 M Tris 1.0 M Tris 2.5 X lO-2 M EDTA 6 % (v/v) acetic acid 1.0 X K H M NaCl 5.0 X lO 2 M EDTA 0.5 % (w/v) SDS (from pH 8 stock) 100 jig / ml proteinase K (Sigma Chemical Co.) Agarose gel (2%) pH 8.0 IX TAE (from 50X stock) Sterilize Tris / EDTA / NaCl 2 % (w/v) agarose by autoclaving and adjust pH (Sigma Chemical Co.) before addition of SDS and 0.05 % (w/v) ethidium proteinase K. bromide (Sigma Chemical Co.) TEN buffer 1.0 X 10-2 M Tris Gel running buffer 1.0 X 10-3 M EDTA IX TAE 5.0 X lO-2 M NaCl 0.05 % (w/v) ethidium Sterilize by autoclaving. bromide

Buffer saturated phenol Gel loading buffer 40 % (v/v) phenol 30 % (v/v) glycerol (Rathburn Chemicals Ltd.) 0.25 % xylene cyanol 12 % (v/v) TEN buffer (Sigma Chemical Co.) 0.04 % (w/v) hydroxyquinoline

PCR reaction buffer (10X RB) 5.0 X 1 0 1 M KC1 2.0 X 1 0 1 M Tris 1.5 X lO 2 M MgCl2.6H20 0.1 % (w/v) gelatin pH 8.4 Autoclave KC1 / Tris and adjust pH before addition of autoclaved MgCl2.6H20 / gelatin.

210 Appendix B: Publications arising out of this thesis

Fox, J.D., Briggs, M., Ward, P. and Tedder, R.S. (1990). HHV6 in salivary glands. Lancet 336: 590-593.

Gopal, M.R., Thompson, B.J., Fox, J., Tedder, R.S. and Honess, R.W. (1990) Detection by PCR of HHV6 and EBV DNA in blood and oropharynx of healthy adults and HIV seropositives. Lancet (letter) 335: 1598-1599.

Fox, J.D., Ward, P., Briggs, M., Irving, W., Stammers, T.G. and Tedder, R.S. (1990). Production of IgM antibody to HHV6 in reactivation and primary infection. Epidemiol. Infect. 104: 289-296.

Coumbe, A., Fox, J.D., Briggs, M., Tedder, R.S. and Berry, C.L. (1990). Cytomegalovirus and human herpesvirus 6 in sudden infant death syndrome: An in situ hybridisation study. Pediatr. Pathol. 10: 483-490.

Briggs, E.M., Fox, J.D. and Tedder, R.S. (1990). Human herpesvirus 6. In Principles and Practice of Clinical Virology, 2nd edition (edited by A.J. Zuckerman, J.E. Banatvala and J.R. Pattison). John Wiley and Sons, Chichester.

Read, R., Larson, E., Harvey, J., Edwards, A., Thomson, B. and Fox, J.D. (1990). Clinical and laboratory findings in the Paul-Bunnell negative Glandular Fever-fatigue Syndrome. J. Infect. 21; 157-165.

Fox, J.D., Briggs, M. and Tedder, R.S. (1988). Antibody to human herpesvirus 6 in HIV-1 positive and negative homosexual men. Lancet (letter) ii: 396-397.

Briggs, M., Fox, J.D. and Tedder, R.S. (1988). Age prevalence of antibody to human herpesvirus 6. Lancet (letter) i: 1058-1059.

In press:

Coyle, P.V., Briggs, M., Tedder, R.S. and Fox, J.D. (1992). A comparison of three immunoassays for the detection of anti-HHV6. J. Virol. Methods.

211 Submitted:

Wakefield, A.J., Fox, J.D., Sawyerr, A.M., Taylor, J.E., Sweenie, C.H., Smith, M., Emery, V., Tedder, R.S. and Pounder, R.E. (1992). Detection of herpesvirus DNA in the large intestine of patients with ulcerative colitis and Crohn's disease using the nested polymerase chain reaction.

In preparation

Fox, J.D., Sweenie, C.H., Briggs, M., Coyle, P.V. and Tedder, R.S. (1992). Reactivation of HHV6 during pregnancy and transmission to newborn offspring by saliva.

Fox, J.D. and Briggs, E.M. (1992). Human herpesviruses 6 and 7. In Principles and Practice of Clinical Virology, 3rd edition (edited by A .J. Zuckerman, J.E. Banatvala and J.R. Pattison). John Wiley and Sons, Chichester.

212 References

Ablashi, D.V., Salahuddin, S.Z., Josephs, S.F., Imam, F., Lusso, P. and Gallo, R.C. (1987). HBLV (or HHV-6) in human cell lines. Nature 329: 207.

Ablashi, D.V., Josephs, S.F., Buchbinder, A., Heilman, K., Nakamura, S., Liana, T., Lusso, P., Kaplan, M., Dahlberg, J., Memon, S., Imam, F., Ablashi, K.L., Markham, P.D., Kramarsky, B.K., Krueger, G.R.F., Biberfeld, P., Wong-Staal, F., Salahuddin, S.Z. and Gallo, R.C. (1988). Human B-lymphotropic virus (human herpesvirus-6). /. Virol. Methods 21: 29-48.

Ablashi, D.V., Salahuddin, S.Z., Streicher, H.Z., Kaplan, M., Krueger, G.R.F., Shearer, G.M. and Gallo, R.C. (1988a). Elevated HBLV (human herpesvirus-6) antibody in HIV-1 antibody-positive symptomatic and asymptomatic individuals. IVth International Conference on AIDS (Stofcholm), Abstract 1134.

Ablashi, D.V., Lusso, P., Hung, C-L., Salahuddin, S.Z., Josephs, S.F., Liana, T., Karmasky, B., Biberfeld, P., Markham, P.D. and Gallo, R.C. (1988b). Utilization of human hematopoietic cell lines for the propagation and characterization of HBLV (human herpesvirus 6). Int. J. Cancer 42: 787-791.

Ablashi, D.V., Zompetta, C., Lease, C., Josephs, S.F., Balachandra, N., Komaroff, A.L., Krueger, G.R.F., Henry, B., Lukau J., and Salahuddin, S.Z. (1991). Human Herpesvirus 6 (HHV6) and chronic fatigue syndrome. Can. Dis. Wkly. Rep. 17: (Suppl.) 33-37.

Adler, S.P. (1983). Transfusion-associated cytomegalovirus infections. Rev. Infect. Dis. 5: 977-993.

Agut, H., Guetard, D., Collandre, H., Dauguet, C., Montagnier, L., Miclea, J., Baurmann, H. and Gessain, A. (1988). Concomitant infection by human herpesvirus 6, HTLV-1 and HIV-2. Lancet i: 712.

Agut, H., Keouedan, D., M'Pele, P., Collandre, H., Samba-Lefevre, C., Rosenheim, M., Ingrand, D., Itoua-Ngaporo, A., Gentilini, M., Montagnier, L. and Huraux, J.M. (1989). Co-cultivation of human herpesvirus 6 and HIV-1 from blood lymphocytes. Res. Virol. 140: 23-26.

Agut, H., Collandre, H., Aubin, J-T., Guetard, D., Favier V., Ingrand, D., Montagnier, L. and Huraux, J-M. (1989a). In vitro sensitivity of human herpesvirus 6 to antiviral drugs. Res. Virol. 140: 219-228.

213 Agut, H., Aubin, J-T., and Huraux, J-M. (1991). Homogenous susceptibility of distinct human herpesvirus 6 strains to antivirals in vitro. J. Infect. Dis 163: 1382-1383

Ahlfors, K., Ivarsson, S-A., Harris, S., Svanberg, L., Holmqvist, R., Lemmark, B. and Theander, G. (1984). Congenital cytomegalovirus infection and disease in Sweden and the relative importance of primary and secondary maternal infections. Scand. J. Infect. Dis. 16: 129-137.

Akesson-Johansson, A., Harmenberg, J., Wahren, B. and Linde, A. (1990). Inhibition of human herpesvirus 6 replication by 9-[4-hydroxy-2-(hydroxymethyl) butyl] guanine (2HM-HBG) and other antiviral compunds. Antimicrob. Agents. Chemother. 34: 2417-2419.

Alford, C.A. and Britt, W. (1990). Cytomegalovirus. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M.S. Hirsch, J.L. Melnick, T.P. Monath and B. Roizman). Raven Press, New York; pp 1981-2010.

Allday, M J. and Crawford, D.H. (1988). Role of epithelium in EBV persistence and pathogenesis of B-cell tumours. Lancet i: 855-857.

Anderson, M.J., Khousam, M.N., Maxwell, D.J., Gould, S.J., Happerfield, L.C. and Smith, W J. (1988). Human parvovirus B19 and hydrops fetalis. Lancet i: 535.

Archard, L.C., Bowles, N.E., Behan, P.O., Bell, E.J. and Doyle, D. (1988). Postviral fatigue syndrome: persistence of enterovirus RNA in muscle and elevated creatine kinase. J. R. Soc. Med. 81: 326-331.

Asada, H., Yalcin, S., Balachandra, K., Higashi, K. and Yamanishi, K. (1989). Establishment of titration system for human herpesvirus 6 and evaluation of neutralizing antibody response to the virus. J. Clin. Microbiol. 27: 2204-2207.

Asano, Y., Yoshikawa, T., Suga, S., Yazaki, T., Hirabayashi, S., Ono, Y., Tsuzuki, K. and Oshima, S. (1989). Human herpesvirus 6 harbouring in kidney. Lancet ii: 1391.

Asano, Y., Yoshikawa, T., Suga, S., Yazaki, T., Hata, T., Nagai, T., Kajita, Y., Ozaki, T. and Yoshida, S. (1989a). Viraemia and neutralizing antibody response in infants with exanthem subitum. J. Pediatr. 114: 535-539.

Asano, Y., Suga, S., Yoshikawa, T., Urisu, A. and Yazaki, T. (1989b). Human herpesvirus type 6 infection (exanthem subitum) without fever. /. Pediatr. 115: 264-265.

214 Asano, Y., Yoshikawa, T., Suga, S., Yazaki, T., Ozaki, T., Saito, Y., Hatano, Y. and Takahashi, M. (1990). Enzyme-linked immunosorbent assay for detection of IgG antibody to human herpesvirus 6. J. Med. Virol. 32: 119-123.

Asano, Y., Yoshikawa, T., Suga, S., Yazaki, T., Kondo, K. and Yamanishi, K. (1990a). Fatal fulminant hepatitis in an infant with human herpesvirus-6 infection. Lancet 335: 862-863.

Asano, Y., Nakashima, T., Yoshikawa, T., Suga, S. and Yazaki, T. (1991). Severity of human herpesvirus-6 viremia and clinical findings in infants with exanthem subitum. /. Pediatr. 118: 891-895.

Asano, Y., Yoshikawa, T., Suga, S., Nakashima, T., Yazaki, T., Fukuda, M., Kojima, S. and Matsuyama, T. (1991a). Reactivation of herpesvirus type 6 in children receiving bone marrow transplants for leukemia. New Engl. J. Med 324: 634-635.

Asano, Y., Yoshikawa, T., Suga, S., Hata, T., Yamazaki, T., and Yazaki, T. (1991b). Simultaneous occurrence of human herpesvirus 6 infection and intussusception in three infants. Pediatr. Infect. Dis. J. 10: 335-336.

Aubin, J-T., Collandre, H., Candotti, D., Ingrand, D., Rouzioux, C., Burgard, M., Richard, S., Huraux, J-M. and Agut, H. (1991). Several groups among human herpesvirus 6 strains can be distinguished by Southern blotting and polymerase chain reaction. J. Clin. Microbiol. 29: 367-372.

Aurelius, E., Johansson, B., Skoldenberg, B., Staland, A. and Forsgren, M. (1991). Rapid diagnosis of herpes simplex encephalitis by nested polymerase chain reaction assay of cerebrospinal fluid. Lancet 337: 189-192.

Ausubel, M., Brent, R., Kingston, R.E., Moore, D.M., Seidman, J.G., Smith, J.A. and Struhl, K. (1989). Current Protocols in Molecular Biology. John Wiley and Sons, Chichester.

Baboonian, C., Venables, P., Maini, R., Kangro, H.O. and Osman, H. (1990). Antibodies to human herpesvirus-6 in Sjogren's syndrome. Arthritis Rheum. 33: 1749-1750.

Baer, R., Bankier, A.T., Biggin, M.D., Deininger, P.L., Farrell, P.J., Gibson, T.J., Hatfull, G., Hudson, G.S., Satchwell, S.C., Seguin, C., Tuffnell, P.S. and Barrell, B.G. (1984). DNA sequence and expression of the B95-8 Epstein-Barr virus genome. Nature 310: 207-211.

215 Baker, C.J. and Edwards, M.S. (1990). Group B streptococcal infections. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J.S. Remington and J.O. Klein). WB Saunders Co, Philadelphia; pp 742-811.

Balachandran, N., Oba, D.E. and Hutt-Fletcher, L.M. (1987). Antigenic cross-reactions among herpes simplex virus types 1 and 2, Epstein-Barr virus and cytomegalovirus. J. Virol 61: 1125-1135.

Balachandran, N., Amelse, R.E., Zhou, W.W. and Chang, C.K. (1989). Identification of proteins specific for human herpesvirus 6-infected human T Cells. /. Virol 63: 2835-2840.

Balachandran, N., Tirawatnapong, S., Pfeiffer, B., Ablashi, D.V., and Salahuddin, S.Z. (1991). Electrophoretic analysis of human herpesvirus 6 polypeptides immunoprecipitated from infected cells with human sera. J. Infect. Dis. 163: 29-34.

Bames, D.M. (1986). Mystery disease at Lake Tahoe challenges virologists and clinicians. Science 234: 541-542.

Baurmann, H., Miclea, J-M., Ferchal, F., Gessain, A., Daniel, M-T., Guetard, D., Collandre, H., Agut, H., Castaigne, S., Rain, J-D., Montagnier, L., Perol, Y., Sigaux, F. and Degos, L. (1988). Adult T-cell leukemia associated with HTLV-1 and simultaneous infection by human immunodeficiency virus type 2 and human herpesvirus 6 in an African woman: A clinical, virologic and familial study. Am. J. Med. 85: 853-857.

Becker, W.B., Engelbrecht, S., Becker, M., Piek, C., Robson, B., Wood, L. and Jacobs, P. (1988). Isolation of a new human herpesvirus producing a lytic infection of helper (CD4) T-lymphocytes in peripheral blood lymphocyte cultures-another cause of acquired immunodeficiency? S. Afr. Med. J. 74: 610-614.

Becker, W.B., Engelbrecht, S., Becker, M., Piek, C., Robson, B., Wood, L. and Jacobs, P. (1989). New T-lymphotropic human herpesviruses. Lancet i: 41.

Bell, E.J., McCartney, R.A. and Riding, M.H. (1988). Coxsackie B viruses and myalgic encephalomyelitis. J. R. Soc. Med. 81: 329-331.

Berenberg, W., Wright, S. and Janeway, C.A. (1949). Roseola infantum (exanthem subitum). New Engl J. Med. 241: 253-259.

Best, J.M., Banatvala, J.E. and Watson, D. (1969). Serum IgM and IgG responses in postnatally acquired rubella. Lancet ii: 65-68.

216 Best, J.M. and O'Shea, S. (1989). Rubella virus. In Diagnostic procedures for Viral, Rickettsial and Chlamydial Infections, 6th edition (edited by N.J.Schmidt and R.W. Emmons). American Public Health Association; pp 731-796.

Biberfeld, P., Kramarsky, B.K., Salahuddin, S.Z. and Gallo, R.C. (1987). Ultrastructural characterization of a new human B lymphotropic DNA virus (human herpesvirus 6) isolated from patients with lymphoproliferative disease. J. Natl. Cancer Inst. 79: 933-941.

Biberfeld, P., Petren, A-L., Eklund, A., Lindemalm, C., Barkhem, T., Ekman, M. Ablashi, D., and Salahuddin, Z. (1988). Human herpesvirus-6 (HHV-6, HBLV) in sarcoidosis and lymphoproliferative disorders. J. Virol. Methods 21: 49-59.

Black, J.B., Sanderlin, K.C., Goldsmith, C.S., Gary, H.E., Lopez, C. and Pellett, P. (1989). Growth properties of human herpesvirus-6 strain Z29. J. Virol. Methods 26: 133-146.

Bland, M. (1987). An Introduction to Medical Statistics. Oxford University Press.

Borisch Chappuis, B., Ellinger, K., Neipel, F., Kirchner, T., Kujath, P., Fleckenstein, B. and Muller-Hermelink, H.K. (1989). Human herpesvirus 6 in lymph nodes. Lancet i: 40.

Bradstreet, C.M. and Taylor, C.E.D. (1962). Technique of complement fixation test applicable to the diagnosis of viral diseases. Month. Bull. Ministry Health Public Health Lab. Service 21: 96-104.

Braun, R.W. and Reiser, H.C. (1986). Replication of human cytomegalovirus in human peripheral blood T cells. J. Virol. 60: 29-36.

Brigati, D.J., Myerson, D., Leary, J.J., Spalholz, B., Travis, S.Z., Fong, C.K.Y., Hsiung, G.D. and Ward, D.C. (1983). Detection of viral genomes in cultured cells and paraffin- embedded tissue sections using biotin-labeled hybridization probes. Virology 126: 32-50.

Brown, N.A., Sumaya, C.V., Liu,C-R., Ench, Y., Kovacs, A., Coronesi, M. and Kaplan, M.H. (1988). Fall in human herpesvirus-6 seropositivity with age. Lancet ii: 396.

Brown, N.A., Kovacs, A. Lui, C-R., Hur, C., Zaia, J.A. and Mosely, J.W. (1988a). Prevalence of antibody to human herpesvirus 6 among blood donors infected with HTV. Lancet ii: 1146.

217 Brytting, M., Sundqvist, V-A., St&lhandske, P., Linde, A. and Wahren, B. (1991). Cytomegalovirus DNA detection of an immediate early protein gene with nested primer oligonucleotides. J. Virol. Methods. 32: 127-138.

Buchbinder, A., Josephs, S.F., Ablashi, D.V., Salahuddin, S.Z., Klotman, M.E., Manak, M., Krueger, G.R.F., Wong-Staal, F. and Gallo, R.C. (1988). Polymerase chain reaction amplification and in situ hybridization for the detection of human B-lymphotropic virus. /. Virol. Methods 21: 191-197.

Buchbinder, A., Ablashi, D.V., Saxinger, C., Josephs, S.F., Salahuddin, S.Z., Gallo, R.C., Biberfeld, P. and A Linde (1989). Human herpesvirus-6 and cross-reactivity with other herpesviruses. Lancet i: 217.

Buchwald, D., and Komaroff, A.L. (1991). Review of laboratory findings for patients with chronic fatigue syndrome. Rev. Infect. Dis. 13: (Suppl. 1) S12-18.

Bums, J., Redfem, D.R.M., Esiri, M.M. and Mcgee, J. O'D. (1986). Human and viral gene detection in routine paraffin embedded tissue by in situ hybridisation with biotinylated probes: viral localisation in herpes encephalitis. J. Clin. Pathol. 39: 1066-1073.

Bums, W.H. and Sandford, G.R. (1990). Susceptibility of human herpesvirus 6 to antivirals in vitro. J. Infect. Dis. 162: 634-637.

Campbell, M.E.M., McCorkindale, S., Everett, R.D. and Onions, D.E. (1991). Activation of gene expression by human herpesvirus 6 is reporter gene-dependent. /. Gen. Virol 72: 1123-1130.

Carrigan, D.R., Knox, K.K. and Tapper, M.A. (1990). Suppression of human immunodeficiency virus type 1 replication by human herpesvirus-6. J. Infect. Dis. 162: 844-851.

Carrigan, D.R., Drobyski, W.R., Russler, S.K., Tapper, M.A., Knox, K.K. and Ash, R.C. (1991). Interstitial pneumonitis associated with human herpesvirus-6 infection after marrow transplantation. Lancet 338: 147-149 .

Ceccherini-Nelli, L., Panicucci, F., Torelli, G., Santini, G., D'Agaro, P. and Bendinelli, M. (1990). Seroepidemiology of HHV6 infection in Italy. VHIth International Congress of Virology (Berlin), Abstract P35-036.

218 Chang, C.K. and Balachandran, N. (1991). Identification, characterization, and sequence analysis of a cDNA encoding a phosphoprotein of human herpesvirus 6. /. Virol 65: 2884-2894.

Chee, M.S., Bankier, A.T., Beck, S., Bohni, R., Brown, C.M., Cemy, R., Horsnell, T., Hutchinson III, C.A., Kouzarides, T., Martigignetti, J.A., Preddie, E., Satchwell, S.C., Tomlinson, P., Weston, K.M. and Barrell, B.G. (1990). Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD 169. Curr. Top. Microbiol. Immunol. 154: 126-169.

Cherry, J.D. (1990). Enteroviruses. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J.S. Remington and J.O. Klein). WB Saunders Co, Philadelphia; pp 325-366.

Chou, S. and Scott, K.M. (1990). Rises in antibody to human herpesvirus 6 detected by enzyme immunoassay in transplant recipients with primary cytomegalovirus infection. J. Clin. Microbiol. 28: 851-854.

Cines, D.B. (1989). Disorders associated with antibodies to endothelial cells. Rev. Infect. Dis. 11: (Suppl. 4) S705-S711.

Clapham, P.R., Weiss, R.A., Dalgleish, A.G., Exley, M., Whitby, D. and Hogg, N. (1987). Human immunodeficiency virus infection of monocytic and T-lymphocytic cells: receptor modulation and differentiation induced by phorbol ester. Virology 158: 44-51.

Clark, D.E., Alexander, F.E., McKinney, P.A., Roberts, B.E. O'Brien, C., Jarrett, R.F., Cartwright, R.A. and Onions, D.E. (1990). The seroepidemiology of human herpesvirus-6 (HHV-6) from a case-control study of leukaemia and lymphoma. Int. J. Cancer 45: 829-833.

Cleary, M.L. and Sklar, J. (1984). Lymphoproliferative disorders in cardiac transplant recipients are multiclonal lymphomas. Lancet ii: 489-493.

Coates, P.J., Hall, P.A., Butler, M.G. and D'Ardenne, A.J. (1987). Rapid technique of DNA-DNA in situ hybridisation on formalin fixed tissue sections using microwave irradiation. J. Clin. Pathol. 40: 865-869.

Coumbe, A., Fox, J.D., Briggs, M., Tedder, R.S. and Berry, C.L. (1990). Cytomegalovirus and human herpesvirus-6 in sudden infant death syndrome: an in situ hybridization study. Pediatr. Pathol. 10: 483-490.

219 Cradock-Watson, J.E., Bourne, M.S. and Vandervelde, E.M. (1972). IgG, IgA and IgM responses in acute rubella determined by the immunofluorescent technique. /. Hyg. Camb. 70: 473-485.

Cradock-Watson, J.E, Ridehalgh, M.K.S and Bourne, M.S. (1979). Specific Immunoglobulin responses after varicella and herpes zoster. /. Hyg. Camb. 82: 319-336.

Cranage, M.P., Kouzarides, T., Bankier, A.T., Satchwell, S., Weston, K., Thomlinson, P., Barrell, B., Hart, H., Bell, S.E., Minson, A.C. and Smith, G.L. (1986). Identification of the human cytomegalovirus glycoprotein B gene and induction of neutralizing antibodies via its expression in recombinant vaccinia virus. EMBO J. 5: 3057-3063.

Crawford, D. and Edwards, J. (1990). Epstein-Barr virus. In Principles and Practice of Clinical Virology, 2nd edition (edited by A.J. Zuckerman, J.E. Banatvala and J. R. Pattison). John Wiley and Sons, Chichester; pp 103-128.

Crawford, D.H., Thomas, J.A., Janossy, G., Sweny, P., Fernando, O.N., Moorhead, J.F. and Thompson, J.H. (1980). Epstein Barr virus nuclear antigen positive lymphoma after cyclosporin A treatment in patient with renal allograft. Lancet i: 1355-1356.

Crouse, C.A., Pflugfelder, S.C., Cleary, T., Demick, S.M. and Atherton, S.S. (1990). Detection of Epstein-Barr virus genomes in normal human lacrimal glands. J. Clin. Microbiol. 28: 1026-1032.

Dahl, H., Linde, A., Sundqvist, V-A. and Wahren, B. (1990). An enzyme-linked immunosorbent assay for IgG antibodies to human herpesvirus 6. J. Virol. Methods 29: 313-324.

Darlington, J., Super, M., Patel, K., Grundy, J.E., Griffiths, P.D. and Emery, V.C. (1991). Use of the polymerase chain reaction to analyse sequence variation within a major neutralizing epitope of glycoprotein B (gp58) in clinical isolates of human cytomegalovirus. J. Gen. Virol. 72: 1985-1989.

Daugherty, S.A., Henry, B.E., Peterson, D.L., Swarts, R.L., Bastien, S. and Thomas, R.S. (1991). Chronic fatigue syndrome in Northern Nevada. Rev. Infect. Dis 13: (Suppl. 1) S39-S44.

Davison, A.J. and Scott, J.E. (1986). The complete DNA sequence of varicella-zoster virus. J. Gen. Virol. 67: 1759-1816.

220 Demmler, G.J., Buffone, G.J., Schimbor, C.M. and May, R.A. (1988). Detection of cytomegalovirus in urine from newborns by using polymerase chain reaction DNA amplification. J. Infect. Dis. 158: 1177-1183.

Di Luca, D., Katsafanas, G., Schirmer, E.C., Balachandran, N. and Frenkel, N. (1990).The replication of viral and cellular DNA in human herpesvirus 6-infected cells. Virology 175: 199-210.

Downing, R.G., Swankambo, N., Serwadda, D., Honess, R., Crawford, D., Jarrett, R. and Griffin, B.E. (1987). Isolation of human lymphotropic herpesviruses from Uganda. Lancet ii: 390.

Dworsky, M., Yow, M., Stagno, S., Pass, R.F. and Alford, C. (1983). Cytomegalovirus infection of breast milk and transmission in infancy. Pediatrics 72: 295-299.

Edson, C.M., Hosier, B.A., Respess, R.A., Waters, D.J. and Thorley-Lawson, D.A. (1985). Cross-reactivity between herpes simplex virus glycoprotein B and a 63,000- Dalton varicella-zoster virus envelope glycoprotein. J. Virol. 56: 333-336.

Efstathiou, S., Gompels, U.A., Craxton, M.A., Honess, R.W. and Ward, K. (1988). DNA homology between a novel human herpesvirus (HHV-6) and human cytomegalovirus. Lancet i: 63-64.

Eizuru, Y., Minematsu, T., Minamishima, Y., Kikuchi, M., Yamanishi, K., Takahashi, M. and Kurata, T. (1989). Human herpesvirus 6 in lymph nodes. Lancet i: 40.

Enders, G., Biber, M., Meyer, G. and Helftenbein, E. (1990). Prevalence of antibodies to human herpesvirus 6 in different age groups, in children with exanthema subitum, other acute exanthematous childhood diseases, Kawasaki syndrome, and acute infections with other herpesviruses and HIV. Infection 18: 12-15.

Ensoli, B., Lusso, P., Schachter, F., Josephs, S.F., Rappaport, J., Negro, F., Gallo, R.C. and Wong-Staal, F. (1989). Human herpesvirus-6 increases HIV-1 expression in co infected T cells via nuclear factors binding to the HIV-1 enhancer. EMBO J. 8: 3019-3027.

Epstein, M.A. (1986). Epstein-Barr virus, infection and tumour induction. Nature 321: 653-654.

Farmer, G.W., Vincent, M.M., Fuccillo, D.A., Horta-Barbosa, L., Ritman, S., Sever, J.L. and Gitnick, G.L. (1973). Viral investigations in ulcerative colitis and regional enteritis. Gastroenterology 65: 8-18.

221 Farr, T.J., Harnett, G.B., Pietroboni, G.R. and Bucens, M.R. (1990). The distribution of antibodies to HHV-6 compared with other herpesviruses in young children. Epidemiol. Infect. 105: 603-607.

Fegan, K.G., Behan, P.O. and Bell, E.J. (1983). Myalgic encephalomyelitis-a report of an epidemic. J. R. Coll. Gen. Pract. 33: 335-337.

Feldman, S., Hughs, W.T. and Daniel, C.B. (1975). Varicella in children with cancer: seventy-seven cases. Pediatrics 56: (Suppl. 4) 388-397.

Fiala, M., Cone, L.A., Chang, C-M. and Mocarski, E.S. (1986). Cytomegalovirus viremia increases with progressive immune deficiency in patients infected with HTLV-III. AIDS Res. Hum. Retroviruses 2: 175-181.

Fingeroth, J.D., Weis, J.J., Tedder, T.F., Strominger, J.L., Biro, P.A. and Fearon, D.T. (1984). Epstein-Barr virus receptor of human B lymphocytes is the C3d receptor CR2. Proc. Natl. Acad. Sci USA 81: 4510-4514.

Flamand, L., Gosselin, J., D'Addario, M., Hiscott, J., Ablashi, D.V., Gallo, R.C. and Menzes, J. (1991). Human herpesvirus 6 induces interleukin-16 and tumour necrosis factor alpha, but not interleukin-6, in peripheral blood mononuclear cell cultures. J. Virol. 65:5105-5110.

Fleisher, G. and Bolognese, R. (1983). Persistent Epstein-Barr virus infection and pregnancy. J. Infect. Dis. 147: 982-986.

Fox, R., Pearson, G. and Vaughan, J.H. (1986). Detection of Epstein-Barr virus- associated antigens and DNA in salivary gland biopsies from patients with Sjogren's syndrome. J. Immunol. 137: 3162-3168.

Fox, R., Chilton, T., Scott, S., Benton, L., Howell, F.V. and Vaughan, J.H. (1987). Potential role of Epstein-Barr vims in Sjogren's syndrome. Rheum. Dis. Clin. North Am. 13: 275-292.

Fox, R. (1988). Epstein-Barr virus and human autoimmune diseases: possibilites and pitfalls. J. Virol. Methods 21: 19-27. de Franchis, R., Cross, N.C.P., Foulkes, N.S. and Cox, T.M. (1988). A potent inhibitor of Taq polymerase copurifies with human genomic DNA. Nucleic Acids Res. 16: 10355.

222 Frenkel, N., Schirmer, E.C., Wyatt, L.S., Katsafanas, G., Roffman, E., Danovich, R.M. and June, C.H. (1990). Isolation of a new herpesvirus from CD4+ T cells. Proc. Natl. Acad. Sci. USA 87: 748-752

Frenkel, N., Schirmer, E.C., Katsafanas, G. and June, C.H. (1990a). T-cell activation is required for efficient replication of human herpesvirus 6. J. Virol. 64: 4598-4602.

Garson, J.A., van den Berghe, J.A. and Kemshead, J.T. (1987). Novel non-isotopic in situ hybridization technique detects small (1 Kb) unique sequences in routinely G-banded human chromosomes: fine mapping of N-myc and p-NGF genes. Nucleic Acids Res. 15: 4761-4770.

Garson, J.A., Tedder, R.S., Briggs, M., Tuke, P., Glazebrook, J.A., Trute, A., Parker, D., Barbara, J.A.J., Contreras M. and Aloysius S. (1990). Detection of hepatitis C viral sequences in blood donations by "nested" polymerase chain reaction and prediction of infectivity. Lancet 335: 1419-1422.

Gehrz, R.C., Christianson, W.R., Linner, K.M., Conroy, M.M., McCue, S.A. and Balfour, H.H. (1981). Cytomegalovirus-specific humoral and cellular immune responses in human pregnancy. J. Infect. Dis. 143: 391-395.

Gelb, L.D. (1990). Varicella-zoster virus. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M.S. Hirsch, J.L. Melnick, T.P. Monath and B. Roizman). Raven Press, New York; pp 2011-2054.

Gershon, A. (1990). Chickenpox, measles and mumps. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J Remington and J.O. Klein). WB Saunders Co, Philadelphia; pp 395-445.

Gilden, D.H., Vafai, A., Shtram, Y., Becker, Y., Devlin, M. and Wellish, M. (1983). Varicella zoster virus DNA in human sensory ganglia. Nature 306: 478-483.

Giraldo, G., Buonaguro, F.M. and Beth-Giraldo, E. (1989). The role of opportunistic viruses in Kaposi's sarcoma (KS) evolution. APMIS 8: (Suppl.) 62-76.

Glenn, J. (1981). Cytomegalovirus infections following renal transplantation. Rev. Infect. Dis. 3: 1151-1178.

Gold, D., Bowden, R., Sixbey, J., Riggs, R., Katon, W.J., Ashley, R., Obrigewitch, R. and Corey, L. (1990). Chronic Fatigue, a prospective clinical and virologic study. JAMA 264: 48-53.

223 Gopal, M.R., Thomson, B.J., Fox, J., Tedder, R.S. and Honess, R.W. (1990). Detection by PCR of HHV6 and EBV DNA in blood and oropharynx of healthy adults and HIV-seropositives. Lancet 335: 1598-1599.

Greenspan, J.S., Greenspan, D., Lennette, E.T., Abrams, D.I., Conant, M.A., Petersen, V. and Freese, U.K. (1985). Replication of Epstein-Barr virus within the epithelial cells of oral'hairy' leukoplakia, an AIDS-associated lesion. New. Engl. J. Med. 313: 1564-1571.

Griffiths, P.D. (1990). Cytomegalovirus. In Principles and Practice of Clinical Virology, 2nd edition (edited by A.J. Zuckerman, J.E. Banatvala and J.R. Pattison). John Wiley and Sons, Chichester; pp 69-102.

Griffiths, P.D. and Baboonian, C. (1984). A prospective study of primary cytomegalovirus infection during pregnancy: final report. Br. J. Obstet. Gynaecol. 91: 307-315.

Griffiths, P.D. and Kangro, H.O. (1984). A user's guide to the indirect solid-phase radioimmunoassay for the detection of cytomegalovirus-specific IgM antibodies. J. Virol. Methods. 8: 271-282.

Griffiths, P.D. Stagno, S., Pass, R.F., Smith, R.J. and Alford, C.A. (1982). Congenital cytomegalovirus infection: Diagnostic and prognostic significance of the detection of specific immunoglobulin M antibodies in cord serum. Pediatrics 69: 544-549.

Grundy, J.E., McKeating, J.A., Ward, P.J., Sanderson, A.R. and Griffiths, P.D. (1987). B2 Microglobulin enhances the infectivity of cytomegalovirus and when bound to the virus enables class IHLA molecules to be used as a virus receptor. /. Gen. Virol. 68: 793-803.

Grundy, J.E., Lui, S.F., Super, M., Berry, N.J., Sweny, P., Fernando, O.N., Moorhead, J. and Griffiths, P.D. (1988). Symptomatic cytomegalovirus infection in seropositive kidney recipients: reinfection with donor virus rather than reactivation of recipient virus. Lancet i: 132-135.

Guerrant, R.L., Cleary, T.G. and Pickering, L.K. (1990). Microorganisms responsible for neonatal diarrhea. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J. S. Remington and J.O. Klein). WB Saunders Co, Philadelphia; pp 901-980.

Gupta, S. and Vayuvegula, B. (1991). A comprehensive immunological analysis in chronic fatigue syndrome. Scand. J. Immunol. 33: 319-327.

224 Gyde, S.N. (1990). Cancer risk in ulcerative colitis. In Inflammatory bowel Diseases, 2nd edition (edited by R.N. Allan, M.R.B. Keighley, J. Alexander-Williams and C.F. Hawkins). Churchill Livingstone, Edinburgh; pp 563-580.

Hall, C.B. (1989). The rash of roses. Arch. Dermatol. 125: 196-198.

Halpin, D., Wessely, S. (1989). VP-1 antigen in chronic postviral fatigue syndrome. Lancet i: 1028-1029.

Handsfield, H.H., Chandler, S.H., Caine, V.A., Meyers, J.D., Corey, L., Medeiros, E. and McDougall, J.K. (1985). Cytomegalovirus infection in sex partners: evidence for sexual transmission. J. Infect. Dis. 151: 344-348.

Hardy, D.A., Arvin, A.M., Yasukawa, L.L., Bronzan, R.N., Lewinsohn, D.M., Hensleigh, P.A. and Prober, C.G. (1990). Use of the polymerase chain reaction for successful identification of asymptomatic genital infection with herpes simplex virus in pregnant women at delivery. J. Infect. Dis. 162: 1031-1035.

Harnett, G.B., Farr, T.J., Pietroboni, G.R. and Bucens, M.R. (1990). Frequent shedding of human herpesvirus 6 in saliva. J. Med. Virol. 30: 128-130. zur Hausen, H. (1982). Human genital cancer: synergism between two virus infections or synergism between a virus infection and initiating events? Lancet ii: 1370-1372. zur Hausen, H., Schulte-Holthausen, H., Klein, G., Henle, W., Henle, G., Clifford, P. and Santesson, L. (1970). EBV DNA in biopsies of Burkitt tumours and anaplastic carcinomas of the nasopharynx. Nature 228: 1056-1058.

Heath, R.B. and Kangro, H.O. (1990). Varicella zoster. In Principles and Practice of Clinical Virology, 2nd edition (edited by A.J. Zuckerman, J.E. Banatvala and J.R. Pattison). John Wiley and Sons, Chichester; pp 43-68.

Heaton, K.W. (1990). Dietary factors in the aetiology of inflammatory bowel disease. In Inflammatory Bowel Diseases, 2nd edition (edited by R.N. Allan, M.R.B. Keighley, J. Alexander-Williams and C.F. Hawkins). Churchill Livingstone, Edinburgh; pp 165-169.

Hendrix, M.G.R, Salimans, M.M.M., van Boven, C.P.A. and Bruggeman, C.A. (1990). High prevalence of latently present cytomegalovirus in arterial walls of patients suffering from grade III artherosclerosis. Am. J. Pathol. 136: 23-28.

Henle, G. and Henle, W. (1966). Immunofluorescence in cells derived from Burkitt's lymphoma. J. Bact. 91: 1248-1256.

225 Henle, G., Lennette, E.T., Alspaugh, M.A. and Henle, W. (1979). Rheumatoid factor as a cause of positive reactions in tests for Epstein-Barr virus-specific IgM antibodies. Clin. Exp. Immunol. 36: 415-422.

Herbst, H., Niedobitek, G., Kneba, M., Hummel, M., Finn, T., Anagnostopoulos, I., Bergholz, M., Krieger, G. and Stein, H. (1990). High incidence of Epstein-Barr virus genomes in Hodgkin's disease. Am. J. Pathol. 137: 13-19.

Higashi, K., Sads, H., Kurata, T., Ishikawa, K-I., Hayami, M., Spriatna, Y., Sutarman and Yamanishi, K. (1989). Presence of antibody to human herpesvirus 6 in monkeys. /. Gen. Virol. 70: 3171-3176.

Higuchi, R. (1989). Simple and rapid preparation of samples for PCR. In PCR Technology, Principles and Applications for DNA Amplification (edited by H.A. Erlich). M Stockton Press; pp 31-37.

Holmberg, S., Lopez, C.A., Byers, R.H., Rutherford, G.W. and Jaffe, H.W. (1989). Human herpesvirus 6 (HHV6) in a cohort of homosexual men. Vth International Conference on AIDS (Montreal), Abstract MAO 46.

Holmes, G.P., Kaplan, J.E., Stewart, J.A., Hunt, B., Pinsky, P.F. and Schonberger, L.B. (1987). A cluster of patients with a chronic mononucleosis-like syndrome. Is Epstein- Barr virus the cause? JAMA 257: 2297-2302.

Holmes, G.P., Kaplan, J.E., Gantz, N.M., Komaroff, A.L., Schonberger, L.B., Straus, S.E., Jones, J.F., Dubois, R.E., Cunningham-Rundles, C., Pahwa, S., Tosato, G., Zegans, L.S., Purtilo, D.T., Brown, N.A., Schooley, R. and Brus, I. (1988). Chronic fatigue syndrome: a working case definition. Ann. Intern. Med. 108: 387-389

Horvat, R.T., Wood, C. and Balachandran, N. (1989). Transactivation of human immunodeficiency virus promoter by human herpesvirus 6. J. Virol. 63: 970-973.

Horvat, R.T., Wood, C., Josephs, S.F. and Balachandran, N. (1991). Transactivation of the human immunodeficiency virus promoter by human herpesvirus 6 (HHV-6) strains GS and Z-29 in primary human T lymphocytes and identification of transactivating HHV-6 (GS) gene fragments. J. Virol. 65: 2895-2902.

Hotchin, N.A. and Crawford, D.H. (1991). The Diagnosis of Epstein-Barr virus- associated disease. In Current Topics in Clinical Virology (edited by P. Morgan- Capner). Public Health Laboratory Services; pp 115-140.

226 Hotchin, N.A., Read, R., Smith, D.G. and Crawford, D.H. (1989). Active Epstein-Barr virus infection in post-viral fatigue syndrome. J. Infect. 18: 143-150.

Hsu, S-M., Raine, L. and Fanger, H. (1981). Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques. A comparison between ABC and unlabelled antibody (PAP) procedures. J. Histochem. Cytochem. 29: 577-580.

Huang, L-M., Lee, C-Y., Lin, K-H., Chuu, W-M., Lee, P-I., Chen, R-L., Chen, J-M., Lin, D-T. (1990). Human herpesvirus-6 associated with fatal haemophagocytic syndrome. Lancet 336: 60.

Huemer, H.P., Larcher, C., Wachter, H. and Dierich, M. (1989). Prevalence of antibodies to human herpesvirus 6 in human immunodeficiency virus 1-seropositive and -negative intravenous drug addicts. J. Infect. Dis. 160: 549-560.

Irving, W.L. and Cunningham, A.L. (1990). Serological diagnosis of infection with human herpesvirus type 6. Br. Med. J. 300: 156-159.

Irving, W.L., Cunningham, A.L., Keogh, A. and Chapman, J.R. (1988). Antibody to both human herpesvirus 6 and cytomegalovirus. Lancet ii: 630-631.

Irving, W.L., Chang, J., Raymond, D.R., Dunstan, R., Grattan-Smith, P. and Cunningham, A.L. (1990). Roseola infantum and other syndromes associated with acute HHV6 infection. Arch. Dis. Child. 65: 1297-1300.

Irving, W.L., Ratnamohan, V.M., Hueston, L.C., Chapman, J.R. and Cunningham, A.L. (1990a). Dual antibody rises to cytomegalovirus and human herpesvirus type 6: frequency of occurrence in CMV infections and evidence for genuine reactivity to both viruses. J. Infect. Dis. 161: 910-916.

Ishiguro, N., Yamada, S., Takahashi, T., Takahashi, Y., Togashi, T., Okuno, T. and Yamanishi, K. (1990). Meningo-encephalitis associated with HHV6 related exanthem subitum. Acta. Paediatr. Scand. 79: 987-989.

Jahan, N., Razzaque, A., Greenspan, J., Conant, M.A. Josephs, S.F., Nakamura, S. and Rosenthal, L.J. (1989). Analysis of human KS biopsies and cloned cell lines for cytomegalovirus, HIV-1, and other selected DNA virus sequences. AIDS Res. Hum. Retroviruses 5: 225-231.

227 Jarrett, R.F., Gledhill, S., Qureshi, F., Crae, S.H., Madhok, R., Brown, I., Evans, I., Krajewski, A., O'Brien, C.J., Cartwright, R.A., Venables, P. and Onions, D.E. (1988). Identification of human herpesvirus 6 specific sequences in 2 patients with non- Hodgkins lymphoma. Leukemia 2: 496-502.

Jarrett, R.F., Gallagher, A., Gledhill, S., Jones, M.D., Teo, I. and Griffin, B.E. (1989). Variation in restriction map of HHV-6 genome. Lancet i: 448-449.

Jarrett, R.F., Clark, D.E., Josephs, S.F., and Onions, D.E. (1990). Detection of human herpesvirus-6 DNA in peripheral blood and saliva. J. Med. Virol. 32: 73-76.

Jewell, D.P. and Snook, J.A. (1990). Immunology of ulcerative colitis and Crohn's disease. In Inflammatory Bowel Diseases, 2nd edition (edited by R.N. Allan, M.R.B. Keighley, J. Alexander-Williams and C.F. Hawkins), Churchill Livingstone, Edinburgh; pp 127-146.

Jiwa, N.M., Van Gemert, G.W., Raap, A.K., Van de Rijke, F.M., Mulder, A., Lens, P.F., Salimans, M.M.M., Zwaan, F.E., Van Dorp, W. and Van der Ploeg, M. (1989). Rapid detection of human cytomegalovirus DNA in peripheral blood leukocytes of viremic transplant recipients by the polymerase chain reaction. Transplantation 48: 72-76.

Jones, M.D. and Griffin, B.E. (1983). Clustered repeat sequences in the genome of Epstein Barr virus. Nucleic. Acids Res. 11: 3919-3937.

Jordan, M.C., Rousseau, W.E., Stewart, J.A., Noble, G.R. and Chin, T.D.Y. (1973). Spontaneous cytomegalovirus mononucleosis. Ann. Intern. Med. 79: 153-160.

Josephs, S.F., Salahuddin, S.Z., Ablashi, D.V., Schachter, F., Wong-Staal, F. and Gallo, R.C. (1986). Genomic analysis of the human B-lymphotropic virus (HBLV). Science 234: 601-603.

Josephs, S.F., Ablashi, D.V., Salahuddin, S.Z., Kramarsky, B.K., Franza, B.R., Pellett, P., Buchbinder, A., Memon, S., Wong-Staal, F. and Gallo, R.C. (1988). Molecular studies of HHV-6. J. Virol. Methods 21: 179-190.

Josephs, S.F., Buchbinder, A., Streicher, H.Z., Ablashi, D.V., Salahuddin, S.Z., Guo, H- G., Wong-Staal, F., Cossman, J., Raffeld, M., Sundeen, J., Levine, P., Biggar, R., Krueger, G.R.F., Fox, R.I. and Gallo, R.C. (1988a). Detection of human B- lymphotropic virus (human herpesvirus 6) sequences in B cell lymphoma tissues of three patients. Leukemia 2: 132-135.

228 Josephs, S.F., Henry, B., Balachandran, N., Strayer, D., Peterson, D., Komaroff, A.L. and Ablashi, D.V., (1991). HHV-6 reactivation in chronic fatigue syndrome. Lancet 337: 1346-1347.

Kangro, H.O. (1980). Evaluation of a radioimmunoassay for IgM-class antibodies against cytomegalovirus. Br. J. Exp. Pathol. 61: 512-520.

Kangro, H.O. (1990). The diagnosis of virus infections. In Topley and Wilson's Principles of Bacteriology, Virology and Immunity, 8th edition (edited by M.T. Parker and L.H. Collier), Edward Arnold, London; pp 228-242.

Kangro, H.O., Pattison, J.R. and Heath, R.B. (1978). The detection of rubella- specific IgM antibodies by radioimmunoassay. Br. J. Exp. Pathol. 59: 577-583.

Kangro, H.O., Ward, A., Argent, S., Heath, R.B., Cradock-Watson, J.E. and Ridehalgh, M.K.S. (1988). Detection of specific IgM in varicella and herpes zoster by antibody- capture radioimmunoassay. Epidemiol. Infect. 101: 187-195.

Kearney, J.F., Radbruch, A., Liesegang, B. and Rajewsky, K. (1979). A new mouse myeloma cell line that has lost immunoglobulin expression but permits the construction of antibody-secreting hybrid cell lines. J. Immunol. 123: 1548-1550.

Keller, R., Peitchel, R., Goldman, J.N. and Goldman, M. (1976). An IgG-Fc receptor induced in cytomegalovirus-infected human fibroblasts. J. Immunol. 116: 772-777.

Kido, S., Kondo, T., Morishima, T., Takahashi, M. and Yamanishi, K. (1990). Detection of human herpesvirus 6 DNA in throat swabs by polymerase chain reaction. J. Med. Virol. 32: 139-142.

Kikuta, H., Lu, H., Matsumoto, S., Josephs, S.F., and Gallo, R.C. (1989). Polymorphism of human herpesvirus 6 DNA from five Japanese patients with exanthem subitum. J. Infect. Dis. 160: 550-551.

Kikuta, H., Lu, H. and Matsumoto, S. (1989a). Susceptibility of human herpesvirus 6 to acyclovir. Lancet ii: 861.

Kikuta, H., Lu, H., Tomizawa, K. and Matsumoto, S. (1990). Enhancement of human herpesvirus 6 replication in adult human lymphocytes by monoclonal antibody to CD3. J. Infect. Dis. 161: 1085-1087.

Kikuta, H., Nakane, A., Lu, H., Taguchi, Y., Minagawa, T. and Matsumoto, S. (1990a). Interferon induction by human herpesvirus 6 in human mononuclear cells. J. Infect. Dis. 162: 35-38.

229 Kikuta, H., Itami, N., Matsumoto, S., Chikaraishi, T. and Togashi, M. (1991). Frequent detection of human herpesvirus 6 DNA in peripheral blood mononuclear cells from kidney transplant patients. J. Infect. Dis. 163: 925.

Kirchesch, H., Mertens, T., Burkhardt, U., Kruppenbacher, J.P., Hoffken, A., Eggers, H.J. (1988). Seroconversion against human herpesvirus-6 (and other herpesviruses) and clinical illness. Lancet ii: 273-274.

Kishi, M., Harada, H., Takahashi, M., Tanaka, A., Hayashi, M., Nonoyama, M., Josephs, S.F., Buchbinder, A., Schachter, F., Ablashi, D.V., Wong-Staal, F., Salahuddin, S.Z. and Gallo, R.C. (1988). A repeat sequence, GGGTTA, is shared by DNA of human herpesvirus 6 and Marek's disease virus. J. Virol. 62: 4824-4827.

Knez, V., Stewart, J.A. and Ziegler, D.W. (1976). Cytomegalovirus specific IgM and IgG response in humans studied by radioimmunoassay. J. Immunol. 117: 2006-2013.

Knipe, D.M. (1989). The role of viral and cellular nuclear proteins in herpes simplex virus replication. Adv. Virus Res. 37: 85-123.

Knowles, W.D. and Gardner, S.D. (1988). High prevalence of antibody to human herpes-6 and seroconversion associated with rash in two infants. Lancet ii: 912-913.

Kohler, G. and Milstein, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256: 495-497.

Komaroff, A.L. (1988). Chronic fatigue syndromes: relationship to chronic viral infections. J. Virol. Methods 21: 3-10.

Komaroff, A.L., Saxinger, C., Buchwald, D., Geiger, A. and Gallo, R.C. (1988). A chronic 'post viral' fatigue syndrome with neurologic features: Serological association with human herpes virus-6 (HHV-6). Clin. Res. 36: 743A.

Kondo, K., Hayakawa, Y., Mori, H., Sato, S., Kondo, T., Takahashi, K., Minamishima, Y., Takahashi, M. and Yamanishi, K. (1990). Detection by polymerase chain reaction amplification of human herpesvirus 6 DNA in peripheral blood of patients with exanthem subitum. J. Clin. Microbiol. 28: 970-974.

Kondo, K., Kondo, T., Okuno, T., Takahashi, M. and Yamanishi, K. (1991). Latent human herpesvirus 6 infection of human monocytes/macrophages. /. Gen. Virol. 72: 1401-1408.

Krueger, G.R.F. and Ablashi, D.V., (1987). Human B-lymphotropic virus in Germany. Lancet ii: 694.

230 Krueger, G.R.F., Koch, B. and Ablashi, D.V., (1987). Persistent fatigue and depression in patient with antibody to human B-lymphotropic virus. Lancet ii: 36.

Krueger, G.R.F., Ablashi, D.V., Salahuddin, S.Z. and Josephs, S.F., (1988). Diagnosis and differential diagnosis of progressive lymphoproliferation and malignant lymphoma in persistent active herpesvirus infection. J. Virol. Methods 21: 255-264.

Krueger, G.R.F., Koch, B., Ramon, A., Ablashi, D.V., Salahuddin, S.Z., Josephs, S.F., Streicher, H.Z., Gallo, R.C. and Habermann, U. (1988a). Antibody prevalence to HBLV (human herpesvirus-6, HHV-6) and suggestive pathogenicity in the general population and in patients with immune deficiency syndromes. J. Virol. Methods 21: 125-131.

Krueger, G.R.F., Wassermann, K., De Clerck, L.S., Stevens, W.J., Boureois, N., Ablashi, D.V., Josephs, S.F., and Balachandran, N. (1990). Latent herpesvirus-6 in salivary and bronchial glands. Lancet 336: 1255-1256.

Kumar, M.L., Nankervis, G.A., Cooper, A.R. and Gold, E. (1984). Postnatally acquired cytomegalovirus infections in infants of CMV-excreting mothers. J. Pediatr. 104: 669-673.

Kusuhara, K., Ueda, K., Okada, K., Miyazaki, C. and Tokugawa, K. (1991). Do second attacks of exanthema subitum result from human herpesvirus 6 reactivation or reinfection? Pediatr. Infect. Dis. J. 10: 468-469.

Kwok, S. and Higuchi, R. (1989). Avoiding false positives with PCR. Nature 239: 237-238.

Kwok, S., Kellogg, D.E., McKinney, N., Spasic, D., Goda, L., Levenson, C. and Sninsky, J.J. (1990). Effects of primer-template mismatches on the polymerase chain reaction: Human immunodeficiency virus type 1 model studies. Nucleic Acids Res. 18: 999-1005.

Langer, P.R., Waldrop, A.A. and Ward, D.C. (1981). Enzymatic synthesis of biotin- labeled polynucleotides: Novel nucleic acid affinity probes. Proc. Natl. Acad. Sci. USA 78: 6633-6637.

Langman, MJ.S. (1990). Epidemiology of inflammatory bowel disease. In Inflammatory Bowel Diseases, 2nd edition (edited by R.N. Allan, M.R.B. Keighley, J. Alexander-Williams and C.F. Hawkins). Churchill Livingstone, Edinburgh; pp 25-33.

Larcher, C., Huemer, H.P., R Margretter and Dierich, M.P. (1988). Serological cross reaction of human herpesvirus 6 with cytomegalovirus. Lancet ii: 963-964.

231 Larcher, C., Huemer, H.P., Honlinger, M., Maroreiter, R. and Dierich, M.P. (1989). Human herpesvirus 6 (HHV-6): incidence in healthy blood donors and in patients with acute infections caused by other herpesviruses. Wien. Klin. Wochenschr. 101: 448-450.

Lasky, L.A. and Dowbenko, D.J. (1984). DNA sequence analysis of the type-common glycoprotein-D genes of herpes simplex virus types 1 and 2. DNA 3: 23-29.

Lawrence, G.L., Chee, M., Craxton, M.A., Gompels, U.A., Honess, R.W. and Barrell, B.G. (1990). Human herpesvirus 6 is closely related to human cytomegalovirus. J. Virol. 64: 287-299.

Levy, J.A., Ferro, F., Lennette, E.T., Oshiro, L. and Poulin, L. (1990). Characterization of a new strain of HHV-6 (HHV-6SF) recovered from the saliva of an HIV-infected individual. Virology 178: 113-121.

Levy, J.A., Ferro, F., Greenspan, D. and Lennette, E.T. (1990a). Frequent isolation of HHV-6 from saliva and high seoprevalence of the virus in the population. Lancet 335: 1047-1050.

Levy, J.A., Landay, A. and Lennette, E.T. (1990b). Human herpesvirus 6 inhibits human immunodeficiency virus type 1 replication in cell culture. J. Clin. Microbiol. 28: 2362-2364.

Lewis, F.A., Griffiths, S., Dunnicliff, R., Wells, M., Dudding, N. and Bird, C.C. (1987). Sensitive in situ hybridisation technique using biotin-streptavidin-polyalkaline phosphatase complex. J. Clin. Pathol. 40: 163-166.

Linde, A., Dahl, H., Wahren, B., Fridell, E., Salahuddin, Z. and Biberfeld, P. (1988). IgG antibodies to human herpesvirus-6 in children and adults both in primary Epstein- Barr virus and cytomegalovirus infections. J. Virol. Methods 21: 117-123.

Linde, A., Fridell, E., Dahl, H., Andersson, J., Biberfeld, P. and Wahren, B. (1990). Effect of primary Epstein-Barr virus infection on human herpesvirus 6, cytomegalovirus, and measles virus immunoglobulin G titers. J. Clin. Microbiol. 28: 211-215.

Littler, E., Lawrence, G.L., Liu, M-Y., Barrell, B.G. and Arrand, J.R. (1990). Identification, cloning and expression of the major capsid protein gene of human herpesvirus 6. J. Virol. 64: 714-722.

232 Longson, M. (1990). Herpes Simplex. In Principles and Practice of Clinical Virology, 2nd edition (edited by A.J. Zuckerman, J.E. Banatvala and J.R. Pattison). John Wiley and Sons, Chichester; pp 3-42.

Loning, T., Milde, K. and Foss, H-D. (1986). In situ hybridization for the detection of cytomegalovirus (CMV) infection. Application of biotinylated CMV-DNA probes on paraffin-embedded specimens. Virchows Arch. 409: 777-790.

Lopez, C. and Honess, R.W. (1990). Human herpesvirus-6. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M. S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; pp 2055-2062.

Lopez, C., Pellett, P., Stewart, J., Goldsmith, C., Sanderlin, K., Black, J., Warfield, D. andFeonno, P. (1988). Characteristics of human herpesvirus-6. J. Infect. Dis. 157: 1271-1273.

Luka, J., Okano, M. and Thiele, G. (1990). Isolation of human herpesvirus-6 from clinical specimens using human fibroblast cultures. J. Clin. Lab. Anal. 4: 483-486.

Luka, J., Pirruccello, S.J. and Kersey, J.H. (1991). HHV-6 genome in T-cell acute lymphoblastic leukaemia. Lancet 338: 1277-1278.

Lusso, P., Salahuddin, S.Z., Ablashi, D.V., Gallo, R.C., Di Marzo Veronese, F., Markham, P.D. (1987). Diverse tropism of human B-lymphotropic virus (human herpesvirus-6). Lancet ii: 743-744.

Lusso, P., Markham, P.D., Tschachler, E., Di Marzo Veronese, F., Salahuddin, S.Z., Ablashi, D.V., Pahwa, S., Krohn, K. and Gallo, R.C. (1988). In vitro cellular tropism of human B-lymphotropic virus (human herpesvirus-6). /. Exp. Med. 167: 1659-1670.

Lusso, P., Ensoli, B., Markham, P.D., Ablashi, D.V., Salahuddin, S.Z., Tschachler, E., Wong-Staal, F. and Gallo, R.C. (1989). Productive dual infection of human CD4+T lymphocytes by HIV-1 and HHV-6. Nature 337: 370-373.

Lusso, P., Gallo, R.C., DeRocco, S.E. and Markham, P.D. (1989a). CD4 is not the membrane receptor for HHV-6. Lancet i: 730.

Lusso, P., Markham, P.D., DeRocco, S.E. and Gallo, R.C. (1990). In vitro susceptibility of T lymphocytes from chimpanzees {Pan troglodytes) to human herpesvirus 6 (HHV- 6): a potential animal model to study the interaction between HHV-6 and human immunodeficiency virus type 1 in vivo. J. Virol. 64: 2751-2758.

233 Lusso, P., De Maria, A., Malnati, M., Lori, F., DeRocco, S.E., Baseler, M. and Gallo, R.C. (1991). Induction of CD4 and susceptibility to HIV-1 infection in human CD8+T lymphocytes by human herpesvirus 6. Nature 349: 533-535.

Lusso, P., Malnati, M., De Maria, A., Balotta, C., DeRocco, S.E., Markham, P.D. and Gallo, R.C. (1991a). Productive infection of CD4+ and CD8+ mature human T cell populations and clones by human herpesvirus 6. J. Immunol. 147: 685-691.

Mahalingam, R., Wellish, M., Wolf, W., Dueland, A.N., Cohrs, R., Vafai, A. and Gilden, D. (1990). Latent varicella-zoster viral DNA in human trigeminal and thoracic ganglia. N. Engl J. Med. 323: 627-631.

Marchette, N.J., Melish, M.E., Hicks, R., Kihara, S., Sam, E. and Ching, D. (1990). Epstein-Barr virus and other herpesvirus infections in Kawasaki syndrome. J. Infect. Dis. 161: 680-684.

Maroulis, G.B., Buckley, R.H. and Younger, J.B. (1971). Serum immunoglobulin concentrations during normal pregnancy. Am. J. Obstet. Gynecol. 109: 971-976.

Marshall, G.S., Gesser, R.M., Yamanishi, K. and Starr, S.E. (1991). Chronic fatigue in children: clinical features, Epstein-Barr virus and human herpesvirus 6 serology and long term follow-up. Pediatr. Infect. Dis. J. 10: 287-290.

Martin, M.E.D., Thomson, B.J., Honess, R.W., Craxton, M.A., Gompels, U.A., Liu, M- Y., Littler, E., Arrand, J.R., Teo, I. and Jones, M.D. (1991). The genome of human herpesvirus 6: maps of unit length and concatemeric genomes for nine restriction endonucleases. J. Gen. Virol. 72: 157-168.

McGoech, D.J., Dalrymple, M.A., Davison, A.J., Dolan, A., Frame, M.C., McNab, D., Perry, L.J., Scott, J.E. and Taylor, P. (1988). The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J. Gen. Virol. 69: 1531-1574.

McKeating, J.A., Grundy, J.E., Varghese, Z. and Griffiths, P.D. (1986). Detection of cytomegalovirus by ELISA in urine samples is inhibited by 132 Microglobulin. J. Med. Virol. 18: 341-348.

McKeating, J.A., Griffiths, P.D. and Weiss, R.A. (1990). HIV susceptibility conferred to human fibroblasts by cytomegalovirus-induced Fc receptor. Nature 343: 659-661.

Meyne, J., Ratliff, R.L. and Moyzis, R.K. (1989). Conservation of the human telomere sequence (TTAGGG)n among vertebrates. Proc. Natl. Acad. Sci. USA 86: 7049-7053.

234 Miller, G. (1990). Epstein-Barr virus. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M. S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; 1921-1958.

Minson, A.C. (1989). Herpesviridae. In Andrewes' Viruses of Vertebrates, 5th edition (edited by J. Porterfield). Balliere Tindall, London; pp 293-332.

Morris, D.J., Littler, E., Jordan, D. and Arrand, J.R. (1988). Antibody responses to human herpesvirus 6 and other herpesviruses. Lancet ii: 1425-1426.

Morris, D.J., Kimpton, C.P. and Corbitt, G. (1989). Persistence of cytomegalovirus in peripheral blood from blood donors. Br. Med. J. 299: 1164-1165.

Morris, D.J., Littler, E., Arrand, J.R., Jordan, D., Mallick, N. and Johnson, R. (1989a). Human herpesvirus 6 infection in renal-transplant recipients. N. Engl. J. Med. 320: 1560-1561.

Mortimer, P.P., Tedder, R.S., Hambling, M.H., Shaft, M.S., Burkhardt, F. and Schilt, U. (1981). Antibody capture immunoassay for anti-rubella IgM. J. Hyg. Camb. 86: 139-153.

Morton, R. (1985). Cytomegalovirus and inflammatory bowel disease. Postgrad. Med. J. 61: 1089-1091

Mullis, K.B. and Faloona, F.A. (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Methods Enzymol. 155: 335-350.

Niederman, J.C., Evans, A.S., Subrahmanyan, L. and McCollum, R.W. (1970). Prevalence, incidence and persistence of EB virus antibody in young adults. N. Engl. J. Med. 282: 361-365.

Niederman, J.C., Lui, C-R., Kaplan, M.H. and Brown, N.A. (1988). Clinical and serological features of human herpesvirus-6 infection in three adults. Lancet ii: 817-819.

Okano, M., Luka, J., Thiele, G.M. Sakiyama, Y., Matsumoto, S. and Purtilo, D.T. (1989). Human herpesvirus 6 infection and Kawasaki disease. /. Clin. Microbiol. 27: 2379-2380.

Okuno, T., Takahashi, K., Balachandra, K., Shiraki, K., Yamanishi, K., Takahashi, M. and Baba, K. (1989). Seroepidemiology of human herpesvirus 6 infection in normal children and adults. J. Clin. Microbiol. 27: 651-653.

235 Okuno, T., Higashi, K., K Shiraki, Yamanishi, K., Takahashi, M., Kokado, J., Ishibashi, M., Takahara, S., Sonoda, T., Tanaka, K., Baba, K., Yabuuchi, H. and Kurata, T. (1990). Human herpesvirus 6 infection in renal transplantation. Transplantation 49: 519-522.

Okuno, T., Sao, H., Asada, H., Shiraki, K., Takahashi, M. and Yamanishi, K. (1990a). Analysis of a glycoprotein of human herpesvirus 6 (HHV-6) using monoclonal antibodies. Virology 176: 625-628.

Olive, D.M., A1 Mufti, S., Simsek, M., Fayez, H. and Nakib, W. (1989). Direct detection of human cytomegalovirus from renal transplant patients following polymerase chain reaction amplification. J. Med. Virol. 29: 232-237.

Olson, N.Y. and Olson, L.C. (1986). Virus-associated hemophagocytic syndrome: relationship to herpes group viruses. Ped. Infect. Dis 5: 369-373.

Pagano, J.S., Kenney, S., Markovitz, D. and Kamine, J. (1988). Epstein-Barr virus and interactions with human retroviruses. J. Virol. Methods 21: 229-239.

Parker, R.E. (1979). Introductory Statistics for Biology, 2nd edition. Edward Arnold Ltd, London.

Pass, R.F., Stagno, S., Dworsky, M.E., Smith, R.J. and Alford, C.A. (1982). Excretion of cytomegalovirus in mothers: observations after delivery of congenitally infected and normal infants. J. Infect. Dis. 146: 1-7.

Pass, R.F. (1985). Epidemiology and transmission of cytomegalovirus. J. Infect. Dis. 152: 243-248.

Pietroboni, G.R., Harnett, G.B., Farr, T.J. and Bucens, M.R. (1988). Human herpesvirus type 6 (HHV-6) and its in vitro effect on human immunodeficiency virus (HIV). J. Clin. Path. 41: 1310-1312.

Pietroboni, G.R., Harnett, G.B., Bucens, M.R. andHoness, R.W. (1988a). Antibody to human herpesvirus 6 in saliva. Lancet i: 1059.

Porter, H.J., Quantrill, A.M. and Fleming, K.A. (1988). B19 parvovirus infection of myocardial cells. Lancet i: 535-536.

Porter-Jordan, K., Rosenberg, E.I., Keiser, J.F., Gross, J.D., Ross, A.M., Nasim, S. and Garrett, C.T. (1990). Nested polymerase chain reaction assay for the detection of cytomegalovirus overcomes false positives caused by contamination with fragmented DNA. J. Med. Virol. 30: 85-91.

236 Portolani, M., Cermelli, C., Pietrosemoli, P., Pecorari, M., Farabegoli, F., Cavazzuti, G.B., Bertolani, M.F. and Micheli, A. (1990). Isolation of HHV-6-related virus from children affected by infectious syndrome. Arch. Virol 110: 143-149.

Powell, K.E., Anderson, N.E., Frith, R.W. and Croxson, M.C. (1990). Non-invasive diagnosis of herpes simplex encephalitis. Lancet 355: 357-358.

Preece, P.M., Pearl, K.N. and Peckham, C.S. (1984). Congenital cytpmegalovirus infection. Arch. Dis. Child. 59: 1120-1126.

Purtilo, D.T., Tatsumi, E., Manolova, G., Manolova, Y., Harada, S., Lipscoumb, H., and Krueger, G. (1985). Epstein-Barr virus as an etiological agent in the pathogenesis of lymphoproliferative and aproliferative diseases in immune deficient patients. Int. Rev. Exp. Pathol. 127: 113-183.

Qavi, H.B., Green, M.T., SeGall, G.K. and Font, R.L. (1989). Demonstration of HIV-1 and HHV-6 in AIDS-associated retinitis. Curr. Eye Res. 8: 379-387.

Quinnan, G.V., Masur, H., Rook, A.H., Armstrong, G., Frederick, W.R., Epstein, J., Manischewitz, J.F., Macher, A..M., Jackson, L., Ames, J., Smith, H.A., Parker, M., Pearson, G.R., Parrillo, J., Mitchell, C. and Straus, S.E. (1984). Herpesvirus infections in the acquired immune deficiency syndrome. JAMA 252: 72-77.

Rando, R.F., Pellett, P., Luciw, P.A., Bohan, C.A. and Srinivasan, A. (1987). Transactivation of human immunodeficiency virus by herpesviruses. Oncogene 1: 13-18.

Ranger, S., Patillaud, S., Denis, F., Himmich, A., Sangare, A., M'Boup, S., Itoua- NGaporo, A., Prince-David, M., Chout, R., Cevallos, R. and Agut, H. (1991). Seroepidemiology of human herpesvirus-6 in pregnant women from different parts of the world. J. Med. Virol 34: 194-198.

Razzaque, A. (1990). Oncogenic potential of human herpesvirus-6 DNA. Oncogene Si 1365-1370.

Read, R., Larson, E., Harvey, J., Edwards, A., Thomson, B., Briggs, M. and Fox, J. (1990). Clinical and laboratory findings in the Paul-Bunnell negative glandular fever- fatigue syndrome. J. Infect. 21: 157-165.

Reedman, B.M. and Klein, G. (1973). Cellular localization of an Epstein-Barr virus (EBV)-associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines. Int. J. Cancer 11: 499-520.

237 Rentrop, R., Knapp, B., Winter, H. and Schweizer, J. (1986). Aminoalkylsilane-treated glass slides as support for in situ hybridization of keratin cDNAs to frozen tissue sections under varying fixation and pretreatment conditions. Histochem. J. 18: 271-276.

Reynolds, D.W., Stagno, S., Hosty, T., Tiller, M. and Alford, C.A. (1973). Maternal cytomegalovirus excretion and perinatal infection. N. Engl. J. Med. 289: 1-5.

Reynolds, D.W., Dean, P.H., Pass, R.F. and Alford, C.A. (1979). Specific cell-mediated immunity in children with congenital and neonatal cytomegalovirus infection and their mothers. J. Infect. Dis. 140: 493-499.

Robert, C., Agut, H., Aubin, J-T, Collandre, H., Ingrand, D., Devillechabrolle, A., LeHoang, P. and Huraux, J-M. (1990). Detection of antibodies to human herpesvirus-6 using immunofluorescence assay. Res. Virol. 141: 545-555.

Roberts, W.H., Sneddon, J.M., Waldman, J. and Stephens, R.E. (1989). Cytomegalovirus infection of gastrointestinal endothelium demonstrated by simultaneous nucleic acid hybridization and immunohistochemistry. Arch. Path. Lab. Med. 113: 461-464.

Roche, J.K. and Huang, E-S. (1977). Viral DNA in inflammatory bowel disease. CMV- bearing cells as a target for immune-mediated enterocytolysis. Gastroenterology 72: 228-233.

Rodier, G., Fox, E., Constantine, N.T. and Abbatte, E.A. (1990). HHV-6 in Djibouti-an epidemiological survey in young adults. Trans. R. Soc. Trop. Med. Hyg. 84: 148-150.

Roffman, E. and Frenkel, N. (1990). Interleukin-2 inhibits the replication of human herpesvirus-6 in mature thymocytes. Virology 175: 591-594.

Roffman, E., Albert, J., Goff, J. and Frenkel, N. (1990). Putative site for the acquisition of human herpesvirus 6 virion tegument. J. Virol. 64: 6308-6313.

Roizman, B. (1990). Herpesviridae: a brief introduction. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M.S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; pp 1787-1793.

Roizman, B. and Sears, A.E. (1987). An inquiry into the mechanisms of herpes simplex virus latency. Annu. Rev. Microbiol. 41: 543-571.

Roizman, B. and Sears, A.E. (1990). Herpes simplex viruses and their replication. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M.S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; pp 1795-1841.

238 Roizman, B., Carmichael, L.E., Deinhardt, F., de-The, G., Nahmias, A.J., Plowright, W., Rapp, F., Sheldrick, P., Takahashi, M. and Wolf, K. (1981). Herpesviridae, definition, provisional nomenclature and taxonomy. Intervirology 16: 201-217.

Rola-Pleszczynski, M., Frenkel, L.D., Fuccillo, D.A., Hensen, S.A., Vincent, M.M., Reynolds, D.W., Stagno, S. and Bellanti, J. (1977). Specific impairment of cell- mediated immunity in mothers of infants with congenital infection due to cytomegalovirus. J. Infect. Dis. 135: 386-391.

Rowley, A.H., Whitley, R.J., Lakeman, F.D. and Wolinsky, S. (1990). Rapid detection of herpes-simplex-virus DNA in cerebrospinal fluid of patients with herpes simplex encephalitis. Lancet 335: 440-441.

Riiger, R., Bomkamm, G.W. and Fleckenstein, B. (1984). Human cytomegalovirus DNA sequences with homologies to the cellular genome. J. Gen. Virol. 65: 1351-1364.

Russler, S.K., Tapper, M.A. and Carrigan, D.R., (1989). Susceptibility of human herpesvirus 6 to acyclovir and ganciclovir. Lancet ii: 382.

Russler, S.K., Tapper, M.A., Knox, K.K., Liepins, A. and Carrigan, D.R. (1991). Pneumonitis associated with coinfection by human herpesvirus 6 and Legionella in an immunocompetent adult. Am. J. Pathol. 138: 1405-1411.

Saiki, R.K. (1989). The design and optimization of PCR. In PCR Technology Principles and Amplifications for DNA Amplification (edited by H.A. Erlich). M Stockton Press; pp 1-22.

Saiki, R.K., Scharf, S J., Faloona, F.A., Mullis, K.B., Horn, G.T., Erlich, H.A. and Amheim, N. (1985). Enzymatic amplification of 6-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230: 1350-1354.

Saiki, R.K., Bugawan, T.L., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1986). Analysis of enzymatically amplified 8-globin and HLA-DQa DNA with allelle-specific oligonucleotide probes. Nature 324: 163-166

Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1988). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239: 487-494.

Salacinski, P., Hope, J., McLean, C., Clement-Jones, V., Sykes, J., Price, J. and Lowry, P.J. (1979). A new simple method which allows theoretical incorporation of radioiodine into proteins and peptides without damage. J. Endocr. 81: 13IP.

239 Salahuddin, S.Z., Ablashi, D.V., Markham, P.D., Josephs, S.F., Sturzenegger, S., Kaplan, M., Halligan, G., Biberfeld, P., Wong-Staal, F. Kramarsky, B. and Gallo, R.C. (1986). Isolation of a new virus, HBLV, in patients with lymphoproliferative disorders. Science 234: 596-601.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning, A Laboratory Manual, 2nd edition. Cold Spring Harbor Laboratory Press.

Saxinger, C., Polesky, H., Eby, N., Grufferman, S., Murphy, R., Tegtmeir, G., Parekh, V., Memon, S. and Hung, C. (1988). Antibody reactivity with HBLY (HHV-6) in US populations. J. Virol. Methods 21: 199-208.

Schachter, J. and Grossman, M. (1990). Chlamydia. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J.S. Remington and J.O. Klein). WB Saunders Co, Philadelphia; pp 464-474.

Schirmer, E.C., Wyatt, L.S., Yamanishi, K., Rodriguez, W.J. and Frenkel, N. (1991). Differentiation between two distinct classes of viruses now classified as human herpesvirus 6 (1991). Proc. Natl. Acad. Sci. USA 88: 5922-5926.

Schmidt, N.J. and Emmons, R.W. (1989). General principles of laboratory diagnostic methods. In Diagnostic Procedures for Viral, Rickettsial and Chlamydial Infections, 6th edition (edited by N.J. Schmidt and R.W. Emmons). American Public Health Association; pp 1-36.

Scott, D.A. and Constantine, N. (1990). Human herpesvirus type 6 and HIV infection in Africa. AIDS 4: 1161-1162.

Shibata, D., Martin, W.J., Appleman, M.D., Causey, D.M., Leedom, J.M. and Amheim, N. (1988). Detection of cytomegalovirus DNA in peripheral blood of patients infected with human immunodeficiency virus. J. Infect. Dis. 158: 1185-1192.

Shiraki, K., Okuno, T., Yamanishi, K. and Takahashi, M. (1982). Polypeptides of varicella-zoster virus (VZV) and immunological relationship of VZV and herpes simplex virus (HSV). J. Gen. Virol. 61: 255-269.

Shiraki, K., Okuno, T., Yamanishi, K. and Takahashi, M. (1989). Virion and nonstructural polypeptides of human herpesvirus-6. Virus Res. 13: 173-178.

240 Simmonds, P., Balfe, P., Peutherer, J.F., Ludlam, C.A., Bishop, J.O. and Leigh Brown, A. (1990). Human immunodeficiency virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol 64: 864-872.

Sixbey, J.W., Nedrud, J.G., Raab-Traub, N., Hanes, R.A. and Pagano, J.S. (1983). Epstein-Barr virus replication in oropharyngeal epithelial cells. N. Engl J. Med. 310: 1225-1230.

Skolnik, P.R., Kosloff, B.R. and Hirsch, M.S. (1988). Bidirectional interactions between human immunodeficiency virus type 1 and cytomegalovirus. J. Infect. Dis. 157: 508-514.

Smith, M.S., Mitchell, J., Corey, L., Gold, D., McCauley, E.A., Glover, D. and Tenover, F.C. (1991). Chronic fatigue in adolescents. Pediatrics 88: 195-202.

Sobue, R., Miyazaki, H., Okamoto, M., Hirano, M., Yoshikawa, T., Suga, S. and Asano, Y. (1991). Fulminant hepatitis in primary human herpesvirus-6 infection. New Engl. J. Med. 324: 1290.

Sokal, R.R. and Rohlf, F.J. (1981). Biometry, the Principles and Practice of Statistics in Biological Research, 2nd edition. W.H. freeman and Company, San Francisco.

Soriano, V., Tor, J., Riezu, I., Civeira, P., Cuende, I. and Prieto, J. (1991). Human herpesvirus 6 in Spain: seroprevalence in blood donors and patients with lymphoproliferative disorders or HIV infection. Vox Sang. 60: 63-64.

Spira, T.J., Bozeman, L.H., Sanderlin, K.C., Warfield, D.T., Feorino, P.M., Holman, R.C., Kaplan, J.E., Fishbein, D.B. and Lopez, C. (1990). Lack of correlation between human herpesvirus-6 infection and the course of human immunodeficiency virus infection. J. Infect. Dis. 161: 567-570.

Stagno, S. (1990). Cytomegalovirus. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J.S. Remington and J.O. Klein) WB Saunders Co, Philadelphia; pp 241-281.

Stagno, S. and Whitley, R.J. (1985). Herpesvirus infections of pregnancy. Part 1: Cytomegalovirus and Epstein-Barr virus infections. N. Engl. J. Med. 313: 1270-1274.

Stagno, S. and Whitley, R.J. (1985a). Herpesvirus infections of pregnancy. Part 11: Herpes simplex virus and Varicella-zoster virus infections. N. Engl. J. Med. 313: 1327-1330.

241 Stagno, S., Reynolds, D.W., Tsiantos, A., Fuccillo, D., Smith, R., Tiller M. and Alford, C.A. (1975). Cervical cytomegalovirus excretion in pregnant and nonpregnant women: suppression in early gestation. J. Infect. Dis. 131: 522-527.

Stagno, S., Reynolds, D.W., Pass, R.F. and Alford, C.A. (1980). Breast milk and the risk of cytomegalovirus infection. N. Engl. J. Med. 302: 1073-1077

Stagno, S., Pass, R.F., Dworsky, M.E., Henderson, R.E., Moore, E.G., Walton, P.D. and Alford, C.A. (1982). Congenital cytomegalovirus infection. N. Engl. J. Med. 306: 945-949.

Stammers, T.G. (1988). Roseola infantum-the neglected exanthem. Practitioner 232: 541-544.

Starr, S.E., Tolpin, M.D., Friedman, H.M., Paucker, K. and Plotkin, S.A. (1979). Impaired cellular immunity to cytomegalovirus in congenitally infected children and their mothers.J. Infect. Dis. 140: 500-505.

Steeper, T.A., Horwitz, C.A., Ablashi, D.V., Salahuddin, S.Z., Saxinger, C., Saltzman, R. and Schwartz, B. (1990). The spectrum of clinical and laboratory findings resulting from human herpesvims-6 (HHV-6) in patients with mononucleosis-like illnesses not resulting from Epstein-Barr virus or cytomegalovirus. AJCP 93: 776-783.

Stem, H., Hannington, G., Booth, J. and Moncrieff, D. (1986). An early marker of fetal infection after primary cytomegalovirus infection in pregnancy. Br. Med. J. 292: 718-720.

Stinski, M. (1990). Cytomegalovirus and its replication. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M. S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; pp 1959-1980.

Straus, S.E. (1988). The chronic mononucleosis syndrome. J. Infect. Dis. 157: 405-412.

Straus, S.E., Reinhold, W., Smith, H.A., Ruyechan, W.T., Henderson, D.K., Blaese, R.M. and Hay, J. (1984). Endonuclease analysis of viral DNA from varicella and subsequent zoster infections in the same patient. N. Engl. J. Med. 311: 1362-1364.

Straus, S.E., Tosato, G., Armstrong, G., Lawley, T., Preble, O.T., Henle, W., Davey, R., Pearson, G., Epstein, J., Brus, I. and Blaese, R.M. (1985). Persisting illness and fatigue in adults with evidence of Epstein-Barr virus infection. Ann. Intern. Med. 102: 7-16.

242 Streicher, H.Z., Hung, C.L., Ablashi, D.V., Heilman, K., Saxinger, C., Fullen, J. and Salahuddin, S.Z. (1988). In vitro inhibition of human herpesvirus-6 by phosphonoformate. J. Virol. Methods 21: 301-304.

Suga, S., Yoshikawa, T., Asano, Y., Yazaki, T. and Hirata, S. (1989). Human herpesvirus-6 infection (exanthem subitum) without rash. Pediatrics 83: 1003-1006.

Suga, S., Yoshikawa, T., Asano, Y., Yazaki, T. and Yoshida, S. (1990). Simultaneous infection with human herpesvirus-6 and measles virus in infants. J. Med. Virol 31: 306-311.

Sutherland, S. and Briggs, J.D. (1983). The detection of antibodies to cytomegalovirus in the sera of renal transplant patients by an IgM antibody capture assay. J. Med. Virol 11: 147-159.

Sutherland, S., Christofinis, G., O'Grady, J. and Williams, R. (1991). A serological investigation of human herpesvirus 6 infections in liver transplant recipients and the detection of cross-reacting antibodies to cytomegalovirus. J. Med. Virol 33: 172-176

Swenson, P.D. and Kaplan, M.H. (1985). Rapid detection of cytomegalovirus in cell culture by indirect immunoperoxidase staining with monoclonal antibody to an early nuclear antigen. J. Clin. Microbiol. 21: 669-673.

Tajiri, N., Nose, O., Baba, K. and Okada, S. (1990). Human herpesvirus-6 infection with liver injury in neonatal hepatitis. Lancet 335: 863.

Takahashi, K., Sonoda, S., Kawakami, K., Miyata, K., Oki, T., Nagata, T., Okuno, T. and Yamanishi, K. (1988). Human herpesvirus 6 and exanthem subitum. Lancet ii: 1463.

Takahashi, K., Sonoda, S., Higashi, K., Kondo, T., Takahashi, H., Takahashi, M. and Yamanishi, K. (1989). Predominant CD4 T-lymphocyte tropism of human herpesvirus 6-related virus. J. Virol 63: 3161-3163.

Tedder, R.S. and Wilson-Croome, R. (1980). Detection by radioimmunoassay of IgM class antibody to hepatitis B core antigen: A comparison of two methods. J. Med. Virol 6: 235-247.

Tedder, R.S., Mortimer, P.P. and Lord, R.B. (1981). Detection of antibody to varicella- zoster virus by competitive and IgM-antibody capture immunoassay. J. Med. Virol 8: 89-101.

243 Tedder, R.S., Yao, J.L. and Anderson, M.J. (1982). The production of monoclonal antibodies to rubella haemagglutinin and their use in antibody-capture assays for rubella-specific IgM. J. Hyg. Camb. 88: 335-350.

Tedder, R.S., Briggs, M., Cameron, C.H., Honess, R., Robertson, D. and Whittle, H. (1987). A novel lymphotropic herpesvirus. Lancet ii: 390-393.

Telenti, A., Marshall, W.F. and Smith, T.F. (1990). Detection of Epstein-Barr virus by polymerase chain reaction. J. Clin. Microbiol. 28: 2187-2190.

Teo, I.A., Griffin, B.E. and Jones, M.D. (1991). Characterization of the DNA polymerase gene of human herpesvirus 6. J. Virol. 65: 4670-4680.

Thayer, W.R. (1990). The aetiology of inflammatory bowel disease: the role of infectious agents. In Inflammatory Bowel Diseases, 2nd edition (edited by R.N. Allan, M.R.B. Keighley, J. Alexander-Williams and C.F. Hawkins). Churchill Livingstone, Edinburgh; pp 147-163.

Thomson, B.J., Efstathiou, S. and Honess, R.W. (1991). Acquisition of the human adeno-associated virus type-2 rep gene by human herpesvirus type-6. Nature 351: 78-80

Thomson, B.J., Martin, M.E.D. and Honess, R.W. (1990). The molecular and biological properties of human herpesvirus 6 (HHV-6). In AIDS and the New Viruses, 1st edition (edited by A. Dalgleish and R. Weiss). Academic Press, London; 163-184.

Toorkey, C.B. and Carrigan, D.R. (1989). Immunohistochemical detection of an immediate early antigen of human cytomegalovirus in normal tissues. J. Infect. Dis. 160: 741-751.

Torelli, G., Marasca, R., Luppi, M., Selleri, L., Ferrari, S., Nami, F., Mariano, M.T., Federico, M., Ceccherini-Nelli, L., Bendinelli, M., Montagnani, G., Montorsi, M. and Tullio, A. (1991). Human herpesvirus-6 in human lymphomas: identification of specific sequences in Hodgkin's lymphoma by polymerase chain reaction. Blood 111 2251-2258.

Tsai, C-H. A., Williams, M.V. and Glaser, R. (1990). A monoclonal antibody that neutralizes Epstein-Barr virus, human cytomegalovirus, human herpesvirus 6, and bacteriophage T4 DNA polymerases. Proc. Natl. Acad. Sci. USA 87: 7963-7967.

Ueda, K., Kusuhara, K., Hirose, M., Okada, K., Miyazaki, C., Tokugawa, K., Nakayama, M. and Yamanishi, K. (1989). Exanthem subitum and antibody to human herpesvirus-6. J. Infect. Dis. 159: 750-752.

244 Variend, S. and Pearse, R.G. (1986). Sudden infant death and cytomegalovirus inclusion disease. J. Clin. Pathol. 39: 383-386.

Wakefield, D., Lloyd, A., Dwyer, J., Salahuddin, S.Z. and Ablashi, D.V., (1988). Human herpesvirus 6 and myalgic encephalomyelitis. Lancet i: 1059.

Wakefield, A.J., Sawyerr, A.M., Dhillion, A.P., Pittilo, R.M., Rowles, P.M., Lewis, A.A.M. and Pounder, R.E. (1989). Pathogenesis of Crohn's disease: multifocal gastrointestinal infarction. Lancet ii: 1057-1062.

Wakefield, A.J., Fox, J.D., Sawyerr, A.M., Taylor, J.H., Sweenie, C.H., Smith, M., Emery, V., Hudson, M., Tedder, R.S. and Pounder, R.E. (1992). Detection of herpesvirus DNA in the large intestine of patients with ulcerative colitis and Crohn's disease using the nested polymerase chain reaction. J. Med. Virol. (submitted).

Ward, K.N., Gray, J.J., and Efstathiou, S. (1989). Brief Report: primary human herpesvirus 6 infection in a patient following liver transplantation from a seropositive donor. /. Med. Virol. 28: 69-72.

Webster, A., Lee, C.A., Cook, D.G., Grundy, J.E., Emery, V.C., Kernoff, P.B.A. and Griffiths, P.D. (1989). Cytomegalovirus infection and the progression towards AIDS in haemophiliacs with human immunodeficiency virus infection. Lancet i: 63-66.

Whitley, R.J. (1990). Herpes simplex viruses. In Virology, 2nd edition (edited by B.N. Fields, D.M. Knipe, R.M. Chanock, M.S. Hirsch, J.L. Melnick, T.B. Monath and B. Roizman). Raven Press, New York; 1843-1887.

Whitley, R.J. (1990a). Herpes simplex virus infections. In Infectious Diseases of the Fetus and Newborn Infant, 3rd edition (edited by J.S. Remington and J.O. Klein) WB Saunders Co, Philadelphia; 282-305.

Wielaard, F., Scherders, J., Dagalinckx, C., Middeldorp, J.M., Sabbe, L.J.M. and Belzen, C.V. (1988). Development of an antibody capture IgM-enzyme linked immunosorbent assay for diagnosis of acute Epstein-Barr virus infections. J. Virol. Methods 21: 105-115.

Wolber, R.A. and Lloyd, R.V. (1988). Cytomegalovirus detection by nonisotopic in situ DNA hybridisation and viral antigen immunostaining using a two-color technique. Hum. Pathol. 19: 736-741.

Wolf, H., zur Hausen, H. and Becker, V. (1973). EB viral genomes in epithelial nasopharyngeal carcinoma cells. Nature New Biol. 244: 245-246.

245 Wolf, H., Haus, M. and Wilmes, E. (1984). Persistence of Epstein-Barr virus in the parotid gland. J. Virol. 51: 795-798.

Wrzos, H., Gibbons, J., Abt, P.L., Gifford, R.P.M. and Yang, H.C. (1990). Human herpesvirus 6 in monocytes of transplant patients. Lancet i: 486-487.

Wyatt, L.S. Balachandran, N. and Frenkel, N. (1990). Variations in the replication and antigenic properties of human herpesvirus 6 strains. J. Infect. Dis. 162: 852-857.

Wyatt, L.S., Rodriguez, W.J., Balachandran, N. and Frenkel, N. (1991). Human herpesvirus 7: Antigenic properties and prevalence in children and adults. /. Virol. 65: 6260-6265.

Yadav, M., Umamaheswari, S. and Ablashi, D.V., (1990). Low prevalence of antibody to human herpesvirus-6 (HHV-6) in Kadazans. Southeast Asian J. Trop. Med. Public Health 21: 259-263.

Yadav, M., Umamaheswari, S. and Ablashi, D. (1991). Antibody reactivity with two strains of human herpesvirus-6 in Malaysians. J. Med. Virol. 33: 236-239.

Yamamoto, M., Black, J.B., Stewart, J.A., Lopez, C. and Pellett, P.E. (1990). Identification of a nucleocapsid protein as a specific serological marker of human herpesvirus 6 infection. J. Clin. Microbiol. 28: 1957-1962.

Yamanishi, K., Okuno, T., Shiraki, K., Takahashi, M., Kondo, T., Asano, Y. and Kurata, T. (1988). Identification of human herpesvirus-6 as a causal agent for exanthem subitum. Lancet i: 1065-1067.

Yanagi, K., Harada, S., Ban, F., Oya, A., Okabe, N. and Tobinai, K. (1990). High prevalence of antibody to human herpesvirus-6 and decrease in titer with increase in age in Japan. J. Infect. Dis. 161: 153-154.

Yoshida, T., Yoshiyama, H., Suzuki, E., Harada, S., Yanagi, K. and Yamamoto, N. (1989). Immune response of patients with exantmema subitum to human herpesvirus type 6 (HHV-6) polypeptides. J. Infect. Dis. 160: 901-902

Yoshida, M., Uno, F., Bai, Z., Yamada, M., Nii, S., Sata, T., Kurata, T., Yamanishi, K. and Takahashi, M. (1989a). Electron microscopic study of a herpes-type virus isolated from an infant with exanthem subitum. Microbiol. Immunol. 33: 147-154.

Yoshida, T., Yoshiyama, H., Suzuki, E. and Yamamoto, N. (1990). Seroreactivity to HHV-6 polypeptides in infants with exanthema subitum and their mothers. Eur. J. Pediatr. 149: 450.

246 Yoshikawa, T., Suga, S., Asano, Y., Yazaki, T., Kodama, H. and Ozaki, T. (1989). Distribution of antibodies to a causative agent of exanthem subitum (human herpesvirus-6) in healthy individuals. Pediatrics 84: 675-677.

Yoshikawa, T., Suga, S., Asano, Y., Yazaki, T. and Ozaki, T. (1990). Neutralizing antibodies to human herpesvirus-6 in healthy individuals. Pediatr. Infect. Dis. J. 9: 589-590.

Yoshikawa, T., Suga, S., Asano, Y., Nakashima, T., Yazaki, T., Sobue, R., Hirano, M., Fukuda, M., Kojima S. and Matsuyama, T. (1991). Human herpesvirus-6 infection in bone marrow transplantation. Blood 78: 1381-1384.

Yoshiyama, H., Suzuki, E., Yoshida, T., Kajii, T. and Yamamoto, N. (1990). Role of human herpesvirus 6 infection in infants with exanthema subitum. Pediatr. Infect. Dis 9: 71-74.

Young, L.S., Clark, D., Sixbey, J.W., Rickinson, A.B. (1986). Epstein-Barr virus receptors on human pharyngeal epithelia. Lancet i: 240-242.

Yousef, G.E., Bell, E.J., Mann, G.F., Murugesan, V., Smith, D.G., McCartney, R.A. and Mowbray, J.F. (1988). Chronic enterovirus infection in patients with postviral fatigue syndrome. Lancet i: 146-150.