The Brief Case: Inherited Chromosomally Integrated Human Herpesvirus 6 (HHV-6) in the Age of Multiplex HHV-6 Testing

CASE

A 4,140-g male infant born via emergency Caesarean section to a healthy primigravida at 41 5/7 weeks’ gestation was admitted to the neonatal intensive care unit for progressive hypoxemic respiratory failure, worsening infiltrates on chest X-rays, and hypotension. The infant was placed on meningitic dosing of ampicillin (300 mg/kg/day intravenously [i.v.] divided every 8 hours [q8h]) and cefepime (150 mg/kg/day i.v. divided q8h) for presumed sepsis. He was placed on extracorporeal membrane oxygenation (ECMO) on day of life (DOL) 2 for worsening hypoxemia and pulmonary hypertension. The infant developed a blanching polymorphous erythematous rash of the trunk and extremities, as well as suspected seizures involving the upper extremities, on DOL 4. An electroencephalogram (EEG) suggested diffuse cerebral dysfunction with focal dysfunction in the right temporal lobe. Empiric acyclovir (60 mg/kg/day i.v. divided q8h) was started. Seizure activity was controlled with phenobarbital therapy. The infant underwent ECMO decannulation on DOL 7 and was extubated to nasal continuous airway pressure (NCPAP) on DOL 9.

The infant was afebrile, weaning on NCPAP, and hemodynamically stable on DOL 11 when he developed increasing leukocytosis (27,300 cells/μl; reference range, 8,000 to 15,000 cells/μl) and elevated C-reactive protein (3.9 mg/dl, reference range 0.0 to 0.9 mg/dl), prompting replacement of cefepime with meropenem (120 mg/kg/day i.v. divided q8h) plus initiation of empirical fluconazole (6 mg/kg/day i.v. divided q24h). A lumbar puncture showed cerebrospinal fluid (CSF) pleocytosis with white blood cells (WBC) of 149 cells/mm³ (reference range, 0 to 15 cells/mm³), red blood cells at 38,000 cells/mm³, slightly low glucose at 33 mg/dl (reference range, 37 to 75 mg/dl), and elevated protein at 109 mg/dl (reference range, 30 to 102 mg/dL). The CSF cell count differential was significant for neutrophil predominance at 65% neutrophils (reference range, 0 to 8%), 2% bands (0%), 17% lymphocytes (10 to 30), 10% monocytes (70 to 90), 5% eosinophils (no reference range), and 1% basophils (no reference range). Multiplex PCR testing of the CSF using the FilmArray meningitis/encephalitis (M/E) panel (BioFire Diagnostics, Salt Lake City, UT) was solely positive for human herpesvirus 6 (HHV-6). Seral blood and urine cultures were negative for bacteria and fungi throughout the hospital course. Serum herpes simplex virus 1 and 2 PCRs were negative. The only other positive microbiological test was a tracheal aspirate from DOL 6 that grew few Escherichia coli bacteria.

Based on the positive HHV-6 PCR from CSF and right temporal lobe seizures, acyclovir was replaced with a planned 21-day course of ganciclovir (6 mg/kg i.v. q12h) for presumed HHV-6 meningitis. On the following day, the infant developed a truncal rash consistent with a viral exanthem.

Quantitative PCR (qPCR) testing for HHV-6 on whole blood on DOL 11 yielded 1.64 million copies/ml, raising the concern for inherited chromosomally integrated HHV-6 (iciHHV-6) rather than active infection. A whole-blood sample collected on DOL 19 was sent to the University of Washington (Seattle, WA) for iciHHV-6 digital droplet PCR (ddPCR) testing, revealing an HHV-6B:human ratio of 1.00 per cell, consistent with a diagnosis of iciHHV-6. Repeat blood HHV-6 qPCR on DOL 33 revealed 2.48 million copies/ml. By that time, the infant had completed 21 days each of ganciclovir and meropenem.

The infant was discharged on DOL 40 with a final diagnosis of culture-negative sepsis and meningitis. Serial neurodevelopmental assessments during the first year of life were normal. Testing of the mother’s hair follicles was positive for iciHHV-6B.

DISCUSSION

This case reveals knowledge gaps in pediatrics and challenges faced by clinical microbiologists. HHV-6 is a virus that is largely unknown or not well understood by clinical neonatologists and pediatricians, although new methods of laboratory testing are increasingly identifying HHV-6 as a potential pathogen. In tandem, clinical microbiologists are kept busy determining the most reliable methods of diagnostic testing. The availability of rapid molecular testing for HHV-6 highlights gaps in understanding of the biology of HHV-6, its significance in infection, and interpretation of laboratory test results. Such knowledge will ultimately impact management of disease.

HHV-6 is a ubiquitous betaherpesvirus that is divided into two species (HHV-6A and HHV-6B) (1). While the natural history of HHV-6A still awaits discovery, HHV-6B is known to infect more than 90% of children by 2 years of age, causing roseola, an infantile eruptive fever that is also known as exanthem subitum or sixth disease (2). Pediatricians have traditionally considered roseola a benign childhood illness. However, it has become apparent in the past decade that, along with another less well-known betaherpesvirus, human herpesvirus 7 (HHV-7), HHV-6B is the leading cause of childhood febrile status epilepticus, a condition associated with an increased risk of both hippocampal injury and subsequent temporal lobe epilepsy (3).

Like other human herpesviruses, HHV-6A and HHV-6B establish latency after primary infection. HHV-6B is often the most common virus detected in the hematopoietic cell transplantation (HCT) patient population, with positivity rates of more than 50%. HHV-6B has been associated with graft-versus-host disease, encephalitis, pneumonitis, hepatitis, and mortality after HCT in pediatric and adult patients, but causality remains to be proven (2). Recent genomic and transcriptomic data mining work in brains from Alzheimer’s patients found surprising associations with both HHV-6A and HHV-7 (4).

HHV-6A and HHV-6B are unique among human herpesviruses in that they can establish life-long latency via integration into the subtelomeric regions of human chromosomes (5). If such an integration occurs in a germ cell, any offspring arising from that germ cell will have HHV-6 in each of their own cells, including their own germ cells (iciHHV-6). Individuals with iciHHV-6 have a 50% chance of passing on the integrated state to their offspring. iciHHV-6 is present in 0.5 to 2% of the population, constituting almost 70 million people worldwide, with the majority carrying iciHHV-6B, similar to the mother/infant pair in our study (6).

The chromosomally integrated HHV-6 virus can excise itself from the chromosome and activate under certain circumstances, such as immune deficiency or pregnancy. Both iciHHV-6-positive and -negative newborns may acquire an active transplacental HHV-6 infection from their mothers. Hall et al. studied 43 congenitally HHV-6 infected infants, of whom 37 (86%) were identified as having iciHHV-6 and six (14%) as having a transplacental infection (7). The infants with transplacental infection had low viral loads in their cord blood and peripheral blood samples, and no HHV-6 DNA detected in their hair follicles that would have been indicative of iciHHV-6. A subsequent study determined that active infection matched the sequence of the iciHHV-6 mother, suggesting the possibility of activation and transmission of the mother’s integrated virus (8). Based on the Hall et al. study, approximately one third of the iciHHV-6 mothers passed the virus on to their infants, with unknown consequences (7).

Since almost everyone has been infected with HHV-6 viruses by the time they reach the childbearing years, the passage of antibodies from the pregnant woman to her fetus is expected to protect her newborn infant against infection with HHV-6. However, pediatricians should be aware that a small percentage of infants born (between 0.25 and 0.33%) will have exposure due to HHV-6 transplacental transmission.

Even though it was unknown if the infant in this case had an active transplacental HHV-6 infection, it was deemed safest to begin treatment with intravenous ganciclovir while awaiting iciHHV-6 testing results, given the FilmArray M/E panel positive result for HHV-6, the clinical findings of an abnormal EEG tracing in the right temporal lobe, and the absence of a definite bacterial pathogen. No antivirals are currently licensed specifically to treat HHV-6. Treatment remains similar to treatment for human cytomegalovirus (HCMV), including the antivirals ganciclovir or valganciclovir, cidofovir, and foscarnet.

HHV-6 has recently been included in new, rapid multiplex PCR panels for meningitis/encephalitis and febrile illness. HHV-6 is revealed to be the second most commonly detected target in evaluation studies of the FilmArray M/E panel after enteroviruses (9). Compared to other targets, establishing HHV-6 infection based solely on detection from CSF can be difficult. Since individuals with iciHHV-6 have the virus in every cell, a positive CSF HHV-6 result may simply be due to lysed cells. Acellular samples cannot discriminate between integrated and actively replicating forms of virus and often require clinical interpretation.

This problem was evident in a recent retrospective evaluation of 15 patients, ranging in age from 27 days to 89 years, with a positive result for HHV-6 on the FilmArray M/E panel (10). The study revealed low probability of true HHV-6 encephalitis on clinical grounds despite HHV-6 plasma viral load detected in seven of the 11 patients tested. The three study patients who were tested for iciHHV-6 by ddPCR were confirmed as positive. Quantitative testing on whole blood, a method more readily available to the clinician, may be indicated in these settings to determine pathogenicity. While low copy numbers derived from whole-blood qPCR rules out iciHHV-6, high viral loads may be seen in both cases of iciHHV-6 and acute infection. Preanalytical variables such as WBC count can greatly affect measured HHV-6 viral loads. However, viral loads of >5.5 log₁₀ viral copies/ml whole blood are strongly suggestive of iciHHV-6 (5). Recognizing iciHHV-6 in plasma samples is more difficult, and viral activation can also yield several logarithms in difference in HHV-6 viral loads. A simpler alternative for confirming iciHHV-6 status is a qualitative PCR test that can be done on a fingernail clipping or hair follicle.

Clinical testing for iciHHV-6 has been chiefly performed using ddPCR testing of whole-blood specimens (11). Whole blood is spun down into serum, and ddPCR testing is performed on both cell pellet and serum fractions. In ddPCR, PCRs are physically partitioned into hundreds of thousands of unique reaction chambers with a water-oil emulsion. For iciHHV-6, this enables accurate absolute quantitation of HHV-6 and a human housekeeping gene such as RPP30, along with comparison of their relative copy numbers. An adjusted relative copy number of 1.00 in the cell pellet indicates that for every copy of HHV-6 measured there are two copies of the housekeeping gene measured, consistent with iciHHV-6. Select cases with two integrated copies of HHV-6 have also been detected. Hospitals without access to ddPCR testing can use a cutoff of >5.5 log₁₀ copies per ml in whole-blood qPCR DNA testing to identify iciHHV-6 individuals (5). In the rare instance when a non-iciHHV-6 patient has a viral load of >5.5 log₁₀, the viral load will be greatly diminished within a week, as these acute phase viral loads typically last for only 1 to 2 days.

From a clinical microbiology standpoint, more work is needed to develop tests to specifically detect cases of active infection in individuals with integrated virus. Such testing could have guided our treating team in deciding the need for ganciclovir therapy. This type of testing would also benefit research aimed at understanding the specific role of HHV-6 activation in disease. Testing for expression of HHV-6A/HHV-6B RNA could be useful, but is not available other than in a few research laboratories (12). Aside from preanalytical specimen stability concerns, HHV-6 RNA detection has been disappointingly associated with lower analytical sensitivity than that of HHV-6 DNA and only is uniformly positive at viral loads greater than 1,000 copies/ml in whole blood (13). Complicating matters, viral RNAs can be packaged within virions for other herpesviruses such as HCMV, and thus the presence of viral RNA may not be indicative of current transcription.

For now, we recommend qPCR testing for HHV-6 on whole blood when the FilmArray M/E panel is positive for HHV-6. Treatment decisions on suspected HHV-6 infection in the setting of iciHHV-6 must be made on a clinical basis. Given the growing number of reports on activation of iciHHV-6, a greater understanding of the basic science and clinical manifestations of HHV-6 is needed to determine the best transcripts and assays that predict clinical outcomes in iciHHV-6 and non-iciHHV-6 patients alike. Such information could ultimately help reduce unnecessary antiviral therapy, an important concept in today’s climate of antibiotic stewardship.

SELF-ASSESSMENT QUESTIONS

1. Approximately what percentage of the population is iciHHV-6 positive?

   • a.
     0.03%
   • b.
     1%
   • c.
     10%
   • d.
     50%

1. Quantitative estimates of HHV-6 viral load in whole blood are NOT affected by which of the following?

   • a.
     WBC count
   • b.
     Viral reactivation
   • c.
     Hemoglobin
   • d.
     iciHHV-6 status

1. What is the preferred diagnostic test for confirmation of iciHHV-6 status?

   • a.
     Plasma quantitative viral PCR
   • b.
     Sequencing of integration sites
   • c.
     Qualitative PCR
   • d.
     Droplet digital PCR

For answers to the self-assessment questions and take-home points, see https://doi.org/10.1128/JCM.02018-18 in this issue.