|Home | About | Journals | Submit | Contact Us | Français|
Human herpesvirus 6 (HHV‐6), a T‐lymphotropic enveloped double stranded DNA virus of the Herpesviridae family, can be divided into two major variants, designated A and B. The B variant is associated with exanthema subitum (roseola infantum, 6th disease), characterised by fever and lymphadenopathy, followed or not by a maculo‐papular rash primarily on the neck and trunk. HHV‐6, however, is also an important opportunistic agent in patients with impaired immune systems. In recent years, increased antibody titres and positive amplification of the viral genome by PCR have been shown in chronic fatigue syndrome, lymphoproliferative diseases, autoimmune thyroiditis, Sjögren syndrome, rheumatoid arthritis, Crohn's disease, and sarcoidosis.1 Complications of primary infection in infancy and childhood and simultaneous occurrence of sudden death or short‐term mortality include pneumonitis, hepatosplenomegaly, fulminant hepatitis, aseptic meningitis, intussusception, thrombocytopenic purpura, fatal haemophagocytic syndrome, and disseminated infection.2,3,4
However, the aetiological contribution of the virus in fatal cases is still debated and the presence of HHV‐6 in a latent status supports the possibility that viral DNA may be amplified in tissue without clinical signs for any of the related diseases. In order to examine the relationship of causality between HHV‐6 infection and mortality in infants and children, we investigated HHV‐6 infection in four children, who died suddenly or shortly after hospital admission.
Case 1 is a female baby (body weight 2580 g, body length 47 cm) born by vaginal delivery following 37 weeks' gestation (Apgar 7/8/9). There was a history of twin pregnancy, with death of the co‐twin in the 16th week of gestation and tocolysis from the 24th week of gestation. Following rupture of the membranes, the amniotic liquid was green. Leucocytosis (40200/µl), thrombocytopenia (14000/µl), and C‐reactive protein of 1.1 mg/dl were found. Antibiotic therapy (penicillin and claforan) was started. She fed orally and became stabile. At the age of 2 months she had bloody vomiting, anaemia, fever and hepatomegaly and splenomegaly. Packed erythrocytes (36 ml) were given. Virology tests were negative, with the exception of HHV‐6 showing immunoglobulin (Ig)M and IgG. The child was transferred to a tertiary hospital.
On physical examination, the baby (body weight 3060 g, about 700 g under the 3rd percentile; body length 53 cm, corresponding to the 3rd percentile) had thrombocytopenia (41000/µl) and haemoglobin of 123 g/d. C‐reactive protein was negative. The child had erythematous maculo‐papular exanthema, diffuse lymph node enlargement, muscular hypotonia, and hepatosplenomegaly (liver 6 cm under the costal arch and spleen 5–6 cm under the costal arch). Liver enzymes were increased (γ‐glutamyltranspeptidase 175 U/l, alanine aminotransferase 127 U/l, aspartate aminotransferase 126 U/l, glutamate dehydrogenase 72.1 U/l). A liver biopsy was performed after normalisation of the platelet level.
Liver histology showed an enlargement of the portal tracts with an increase of the connective tissue, but few inflammatory cells. Neither giant cells nor bile pigments were seen. In Kupfer cells haemosiderin pigment was detected. Liver iron content was increased (2930 μg/g; normal 700 μg/g) and bone marrow biopsy showed dysgranulopoiesis. Haemophagocytosis, lymphophagocytosis and foamy macrophages characteristic of a metabolic disease or immunodeficiency disorders were not observed. HLA typing was negative for neonatal haemochromatosis (child: HLA A30, A32, B14, B41, DR1, DR13; mother: HLA A3, A30, B13, B41, Cw6, Bw4, Bw6). Lymphotoxic antibodies (Terasaki) were not found. The clotting system was accurately investigated and IgG‐loaded glycoprotein complexes IIb/IIIa and Ib/IX were found within the normal range.
HHV‐6 DNA was found in the blood, liver tissue (liver biopsy), bone marrow (bone marrow biopsy), and cerebrospinal fluid (lumbar puncture). No toxic antibodies against lymphocytes were found. Total B‐lymphocytes were increased (1838/µl, 39.5%), total T‐lymphocytes and relative T‐cell subpopulations were markedly decreased (T‐cells: 1079/µl (23.2%), T4/T8 ratio 1, killer‐cells 161/µl (3.5%)). IgG values were low (186 mg/dl), but IgM and IgA were normal. The histological examination of a skin biopsy specimen showed acanthosis, focal parakeratosis, and apparent lack of the stratum granulosum, as well infiltration of macrophages, T lymphocytes, and deposition of complement components (C9, C5 and C3d). A graft‐versus‐host‐reaction (GvHR) was excluded because of the lack of linear IgM‐depositions at level of the epidermal‐basal membrane zone), a finding in 40% (acute GvHR) and 85% (chronic GvHR) of GvHR. Focal zones of loss of the corneal layer indicate the diagnosis of psoriasiform epithelial hyperplasia with superficial lichenoid dermatitis. Serum markers showed IgG against herpes simplex virus, cytomegalovirus, Epstein‐Barr virus, rubella virus, parvovirus B19, adenovirus. Human immunodeficiency virus, hepatitis A virus were negative, hepatitis C virus, and hepatitis G virus were negative IgM was always negative. Treponema pallidum infection was also excluded. Toxoplasmosis IgG was 1:64, whereas IgM was negative.
Chest x ray, upper barium enema, skull ultrasound, metabolic status (amino acids and organic acids, mucopolysaccharides, oligosaccharides), and ophthalmic examination were normal. Echocardiography identified an atrial small defect of type II. At the age of 4 months (5 days before death), she developed bowel gangrene and underwent surgery, but died of septic multi‐organ failure with a clinical diagnosis of persisting HHV‐6 infection. Thus, HHV‐6 DNA had been found in the blood, liver tissue, bone marrow, and cerebrospinal fluid. At autopsy, diffuse alveolar damage of the lung, portal fibrosis of the liver, marked involution of the lymphatic organs (thymus, palatinal tonsils, white pulp of the spleen, lymph nodes), ischaemic degeneration of the renal tubules, necrotising enteritis, and skin candidiasis were found. Immunohistochemistry with antibodies against B and T lymphocytes (UCHL1, L26) showed no major immunodeficiency. Permission to perform brain autopsy was not granted. HHV‐6 involvement into death was suggested from the amplification of the viral DNA of HHV‐6 in blood, liver tissue, bone marrow, and cerebrospinal fluid.
Case 2 is a 3‐year‐old male infant that had apparently had exanthema subitum at the age of 2 years. However, about 6 months later, the infant fell ill with rash and temperatures up to 39.7°C. His response to paracetamol was good and prompt, but one day after the onset of symptoms, he was found dead in his cot. The coroner ordered a forensic autopsy. At autopsy, examination of the upper airways showed tracheitis on gross examination. Histological examination of the upper airways confirmed macroscopic diagnosis and added bronchitis. Trachea and bronchi showed a very few diffuse mononuclear cell infiltrates. Brain examination showed marked brain oedema with scattered round cell infiltrates. Thymus showed intraparenchymal haemorrhage. Complete virology tests (serology, cell culture, PCR) excluded infection with herpes simplex virus, varicella zoster virus, cytomegalovirus, influenza A virus, influenza B virus, adenovirus, enterovirus, Mycoplasma sp, and Chlamydia sp. HHV‐6 involvement into death was suggested in this case clinically. HHV‐6 DNA was found in tissue specimens from heart, liver and lung.
Case 3 is a 6‐year‐old male child that was growing up well until 1 month before death. After an infection of the upper airways, he developed dyspnoea, cardiomegaly and pericardial effusion, and was admitted to a local hospital. In consequence of a maculo‐papular rash, HHV‐6 infection was suspected and preventive antibiotic therapy to avoid bacterial complications was started. He developed low cardiac output signs with liver and kidney failure and required high doses of catecholamines and atropine, suggesting the necessity for urgent mitral valve replacement. The child was transferred to a tertiary hospital for cardiac surgery, but died postoperatively the following day. At autopsy, the child (body weight 20 kg; body length 108 cm) had tooth caries; otherwise the mitral valve was correctly replaced without vegetations and the organs showed signs of global heart failure. The heart showed a dilatation of all chambers with hypertrophy of the outflow tracts of both ventricles (heart weight 150 g; heart weight of a normal 6‐year‐old child 94 g). The thickness of the outflow tract of the ventricle was 8 mm and 5 mm at left and right, respectively. Histological examination of the heart showed a diffuse myocarditis with granulomas of rheumatic type with Aschofs nodules and Anitschkow cells. Brain examination showed symmetrical hemispheres with intact circulus arteriosus Willisii and hyperaemia of the leptomengeal blood vessels. The only neuropathological finding was moderately marked brain oedema. HHV‐6 involvement into death was suggested clinically and by laboratory serum analysis.
Case 4 was an 11‐year‐old female child who had profuse diarrhoea during the night and was admitted to the paediatric intensive care unit (PICU) the next morning as she showed respiratory insufficiency and myocardial failure. The child died after 70 min of maximal effort in the PICU, with a clinical diagnosis of rapid viral infection with involvement of the heart. At autopsy, we found myocarditis, a few mononuclear infiltrates in the peribronchial tissue of the lungs and in the portal tracts of the liver, and diffuse enteritis. The initial clinical diagnosis of viral infection was supported from the detection of HHV‐6 DNA in tissue specimens from the heart, small and large bowel, and mesenteric lymph nodes. No infection with varicella zoster virus or enterovirus was found. HHV‐6 involvement into death was suggested at first pathology examination of the postmortem findings. Neuropathology examination showed symmetrical hemispheres without necrosis or haemorrhage. The circulus arteriosus cerebri was normally conformed and leptomeningeal blood vessels showed hyperaemia. There were moderately marked parahippocampal furrows and marked oedema with tigrolysis and cytoplasm swelling of the neurons.
Tissue specimens were collected at rapid autopsy performed 12–24 hours after death (bodies had been maintained at a temperature of 0°C up to the time of autopsy). An indirect immunofluorescence assay was used to detect HHV‐6 antibodies in serum. We used standard procedures according to the manufacturer's recommendations (Viramed, Munich, Germany). Briefly, slides were covered with a mixture of cells infected with HHV‐6 stain U1102 or mock‐infected HSB‐2 cells and non‐infected cells. Sera were diluted 1:20 for detection of IgG antibodies and 1:10 for detection of IgM antibodies. FITC‐labelled goat anti‐human IgG and anti‐human IgM were used. Titres 1:20 for IgG and 1:10 for IgM were regarded as positive.
Conventional HHV‐6 DNA amplification was performed as follows. Purified DNA was extracted from liver, myocardium, lymph nodes, bone marrow, lymphocytes and serum according to the QIAamp tissue kit protocol (QIAGEN, Hilden, Germany). Lymphocytes were purified from whole blood treated with EDTA by ficoll gradient centrifugation. DNA from bone marrow, lymphocytes and serum was purified using the QIAamp blood kit. To detect HHV‐6 DNA, two successive sets of 35 cycles of hot start amplification were performed in a Gene Amp PCR System 2400 thermocycler (Perkin Elmer, Norwalk, CT, USA). Primers for both variants A and B of HHV‐6 were synthesised on a Beckman Oligo 1000 M oligonucleotide synthesiser. In the first round of amplification (50 s at 94°C, 30 s at 50°C and 60 s at 72°C with an additional time extension delay of 10 min at 72°C), outer primers P1 (5′ GTG GCG TTA AGA CGG ATT GT 3′) and P2 (5′ ATC TAT CCC TCG ACT GCT TC 3′) were used to amplify a DNA fragment with a size of 780 base pairs (bp) (16). Each 100 μl of reaction mix included 10 μl extracted DNA and 90 μl master mix (10 mM Tris‐HCl (pH 8.3), 50 mM KCl, 3 mM MgCl2, 200 μM of each dNTP, 0.2 μM of each primer and 2.5 units Taq DNA polymerase (Perkin Elmer). A nested PCR was carried out with 2 μl aliquot of the first PCR reaction, PCR master mix, and inner primers P3 (5′ GTG GGT GTG TTC GGT GTA ACT TTT ATG 3′) and P4 (5′ TTC GAG ATT TAC AAT GCA AGT CTG CGG 3′) which amplified a 626 bp target. Reaction conditions and thermal cycling were as above except for the annealing step that was performed at 52°C. PCR products were electrophoresed in a 2% agarose gel containing 0.5 μg/ml ethidium bromide and visualised under ultraviolet illumination. All biopsy tissues were also extensively evaluated to exclude other human herpesviruses.
RNA detection was carried out by RT‐PCR. Total cell RNA was transcribed into single‐strand cDNA by using Superscript mouse tumour virus reverse transcriptase (GIBCO/BRL, Eggenstein, Germany) according to the supplier's protocol. Single strands of cDNA were phenol extracted twice and precipitated, and 150–250 ng of each was subjected to PCR. The DNA was amplified in a total volume of 100 μl of 2.5 U of Taq polymerase, 100 μM deoxynucleoside triphosphate, 3.75 mM MgCl2, 10 mM Tris‐HCl (pH 8.3), 50 mM KCl, 0.001% gelatine, and 50 pmol of each primer. For amplification of the IE‐1 transcript, primers Kpn 28 (5′ AGA GAG TCT CAT GTG TGA TAC ATC 3′) and 3′‐primer 3′ NTermSac (5′ CAT TGT TCA TAT GAG CTC TCA AAT CC 3′) were used. Transcripts extending into the IE2hom‐region were amplified using oligonucleotide Kpn28 and 3′‐primer IE2Race3 (CAT CTT TTG TAT GGC TGT TG). Amplification was performed for 35 cycles (initial denaturation: 5 min 94°C, followed by 30 s at 94°C, 30 s at 55°C, 1 min at 72°C). PCR products were resolved on 2% agarose gels. All tissues were coded so that the scientists who performed the PCR procedures were unaware of the identity of any patient. All samples were tested in duplicate and in parallel with negative and positive control samples (distilled water, “housekeeping” genes).
Immunohistochemistry for viral detection was carried out on formalin fixed and paraffin embedded tissue according to supplier's instructions (BioTrend, Cologne, Germany). We used mouse monoclonal antibodies (IgG) against the 101 KDa early antigen of HHV‐6 B using a dilution 1:50 and both avidin–biotin complex and catalysed signal amplification methods as immunohistochemistry detection techniques. An in situ hybridisation study was also performed, but no distinct signal was evident. According to the supplier's instructions there is no cross‐reactivity with HHV‐6 A variant. This antibody reacts weakly with cells infected with human cytomegalovirus in immunofluorescence.
Ultrastructural analysis used 1 mm3 of tissue that was fixed in 4% paraformaldehyde/2.5% cacodylate‐buffered cacodylate (Karnovski). The tissue specimen was then postfixed in 1% osmium tetroxide, dehydrated in graded alcohols, and embedded in epoxy resin (Spurr). Ultrathin sections were cut and stained with uranyl acetate and lead citrate after selection of appropriate blocks by examination of the semithin sections. The tissue sections were then examined using a transmission electron microscope EM 910 (LEO, Germany) at 80 kV.
Case 1 showed positive IgM and IgG (maternal) in serum (five times) and was positive for HHV‐6 DNA in serum, liver, lymphocytes and bone marrow by conventional nested PCR. RT‐PCR was, however, negative. Immunohistochemistry of the lung (left upper lobe) was unambiguously positive for HHV‐6 (fig 1A1A).). Immunohistochemistry was, however, equivocally positive in the myocardial tissue, but definitely negative in thymus, liver and skin specimens. Ultrastructural examinations did not show any virus. Case 2 showed amplification of DNA sequences in the lung, liver, and heart, but not in serum, cerebrospinal fluid, brain and intestinal content. Immunohistochemistry of the lung was only weakly positive for HHV‐6. RT‐PCR and ultrastructural examinations were, however, negative. Case 3 had conventional nested PCR positive for myocardium and thoracic lymph nodes. Immunohistochemistry of the myocardium was not clear‐cut positive for HHV‐6. Postmortem serology was negative for IgM, whereas PCR, RT‐PCR and ultrastructural examinations were always negative. Case 4 showed equivocally positive cells in the myocardial tissue, even when a negative control was used (fig 1B–C). Although different techniques were applied to reduce background staining, some staining was seen. Thus, a comprehensive immunohistochemistry study of the myocardium showed focal areas with positive cells for HHV‐6. Positive results were observed by PCR in specimens coming from heart, large bowel, small bowel, and lymph nodes. Conversely, RT‐PCR and ultrastructural examinations were always negative.
Thus, in all four cases, both transcriptosome analysis (fig 1D1D)) and transmission electron microscopy were negative, despite partial positive immunohistochemistry findings and amplification of viral DNA sequences.
One of the central questions concerning nucleic acid amplification and infectious disease in critically ill infants and children admitted to the intensive care unit is whether a positive PCR result can effectively indicate an active infection. In four infants and children, we amplified DNA sequences of HHV‐6 from different tissues, but ultrastructural examination and RT‐PCR failed to identify mRNA sequences of HHV‐6.
The role of active HHV‐6 infection in young children has been widely debated. Very few reports have indicated a possible aetiological role of HHV‐6 in sudden death of infants and children. It has been postulated that HHV‐6 infection might be the cause of both immunodeficiency and fatal opportunistic infection.5 However, no systematic evaluation of effective virus reactivation has been published. There are three initial questions concerning our results that should be discussed. First, we need to establish the probable cause for the positive result using HHV‐6 DNA detection. We think this might be the detection of the latent virus in T lymphocytes, which may be present in low copy numbers. This is a common feature of infections with herpesviruses. Moreover, the virus may be present in the serum after lysis of a few cells, in which it was present in a latent status (see below). A second point is to try to determine the cause for the negative results of RT‐PCR detection. HHV‐6 RNA was not detectable by RT‐PCR, probably due to the fact that the virus is transcriptionally not active and might play a role as innocent bystander. This assumption is also supported by the negative electron microscopic results, where an acute viral infection was also not detectable. Finally, it is necessary to explain the results of the immunohistochemistry, considering the results of RT‐PCR. We consider that the immunohistochemical signal, which is either strong or weak, is similar to the false positive findings seen in peripheral blood mononuclear cells (PBMC) with virus genome integration. Thus, immunohistochemistry can easily detect the incorporation of the viral genome into the human genome and, in our opinion, does not allow a clear response to the aetiology question.
HHV‐6 infections are also a major concern in immunocompromised patients in intensive care units. Certain viruses, such as the family of the Herpesviridae, may reactivate in immunosuppressed patients, making a positive PCR difficult to interpret. RT‐PCR was performed using primer sets in the U42 gene of each viral genome. In saliva samples from 29 healthy adults, HHV‐6 and HHV‐7 DNA were detected in 41.4% and 89.7%, respectively. The average copy number of the HHV‐7 genome in the positive samples was higher than that of the HHV‐6 genome. Follow‐up studies of six seropositive individuals showed that the amount of HHV‐7 DNA was constant in each individual and that “high producers” and “low producers” could be distinguished. The amount of HHV‐6 DNA varied drastically over time in each individual.6 This raises the question of which virological marker should be used for the diagnosis and monitoring of active HHV‐6 infections. We reviewed all methods for diagnosing HHV‐6, their advantages, disadvantages and limitations and found transcriptosome analysis to be the most feasible assay in clinical pathology (table 11).). Further, it has been asserted that the viral load in plasma, as determined by RT‐PCR, is a good indicator for this purpose.7 Indeed, the preparation of samples in the case of plasma‐based assays is easier than that for assays using PBMC. It has been argued that PBMC‐based assays may reflect both latent and active viral infection, but, in a previous study, HHV‐6 DNA loads in PBMC, also measured by real‐time PCR, were usually significantly higher in persons with delayed neutrophil engraftment or severe graft‐versus‐host disease.8 HHV‐6 DNA in plasma is highly sensitive to the action of DNase I, has a detection rate lower than that in PBMCs, and exhibits a distribution in plasma sub‐fractions that parallels that of cellular DNA. The presence of HHV‐6 DNA in plasma may result from the passive lysis of infected blood cells, rather than the active production of mature viral particles. Accordingly, serum or plasma from persons with HHV‐6 sequences integrated in their cellular chromosomes may have high viral DNA loads, even when the subjects are healthy.9 High amounts of HHV‐6 DNA in plasma would mainly reflect a high intracellular viral load. The presence of this marker might be delayed, compared with the initiation of active infection, and exposed to artefacts related to the variability of cell lysis events in clinical specimens.
Our work is important because it investigates the specificity of HHV‐6 as aetiological factor for cause of death in four children with initial diagnosis of HHV‐6 infection. It suggests caution in implicating HHV‐6 in the cause of death. The issue of how immunocompetency is present in a ill child is a good one, but it is difficult to answer. In an attempt to answer this question we thoroughly re‐examined both clinical data and anatomo‐pathological findings including mucosa‐associated lymphatic tissue of our four cases. To the best of our knowledge, patients were immunocompetent, or at least did not show any commonly diagnosable immunodeficiency. However, this affirmation is based on both clinical and pathological data (T‐ and B‐cell markers by immunohistochemistry) for patients who died after hospital admission and on pathological data only (T‐ and B‐cell markers by immunohistochemistry) for the infant with cot death.
It cannot be excluded that HHV‐6 may cause a permanent latent infection in early childhood, but, in our opinion and in consideration of the transcriptosome analysis results, there is no convincing evidence for a lethal outcome. Probably, HHV‐6 may be able to function as a synergistic cofactor in lung infections by other pathogens. Similar to other herpesviruses, HHV‐6 can be reactivated at any time and may cause lung disease if host defence mechanisms become defective. However, transcriptosome and ultrastructural analysis should be able to exclude its reactivation even if a morphological finding would indicate the contrary. This has great importance for quality assurance and standard policies in intensive care units aiming to support better diagnostic procedures for the future.
Thus, in consideration of our data, we question a participation of HHV‐6 in the deaths of our patients and strongly suggest performing transcriptosome analysis and/or transmission electron microscopy before making this diagnosis. It can be judged a quite restrictive procedure, but our data suggest a likely uncertainty of diagnosis with immunohistochemistry, in situ hybridisation or PCR only. We agree that every effort must be made to set up promising approaches for monitoring HHV‐6 infection in immunocompromised patients. The only suitable approach for fully solving the above raised questions should involve clinical studies to compare the abilities of whole‐blood and plasma assays to specifically and accurately detect HHV‐6 infection as early as possible.
We thank Professor Dr G van Kaick, German Cancer Research Institute, Heidelberg, Germany, for iron measurement of liver tissue (case 1); the Institute of Clinical Immunology and Medical Transfusion, University of Giessen, Germany, for the data provided about HHV‐6; Dr F Neipel for transcriptosome primer sequences and RT‐PCR conditions; and Mrs G Gmeiner (Central Medical Institute of German Federal Armed Forces, Koblenz, Germany) for expert technical assistance. Permission was given for these case reports.
This study was partly funded by the Peripro charity and partly by intramural financial support (Medical University of Innsbruck). No private companies were involved in this study.
Competing interests: None.