Abstract
Human herpesviruses may cause infections of the central nervous system during primary infection or following reactivation from a latent state. Especially in immunosuppressed patients the infection can take a life-threatening course, and therefore early diagnosis of herpesvirus-associated neurological diseases should have high priority. Clinical presentation in these patients is usually without typical features, making diagnosis even more challenging. Therefore general broad testing for different herpesviruses in cerebrospinal fluid samples is highly recommended. In addition, determination of the virus DNA level in the cerebrospinal fluid by quantitative assays seems to be of high importance to determine prognosis. Moreover, it might help to differentiate between specific virus-associated disease and unspecific presence of virus in the cerebrospinal fluid, especially in immunocompromised patients. Polymerase chain reaction analysis of cerebrospinal fluid has revolutionized the diagnosis of nervous system viral infections, particularly those caused by human herpesviruses. This review summarizes the role human herpesviruses play in central nervous system infections in immunocompromised patients, with a focus on the clinical manifestation of encephalitis.
Keywords
Introduction
All herpesviruses have a similar molecular structure – they contain double-stranded DNA. There are eight human herpesviruses (HHVs), herpes simplex virus type 1 (HSV-1) and type 2 (HSV-2), varicella-zoster virus (VZV), Epstein-Barr virus (EBV or HHV-4), cytomegalovirus (CMV or HHV-5), HHV-6, HHV-7 and HHV-8 [Kaposi sarcoma (KS) herpesvirus]. These viruses share similar biological characteristics, including the unique ability to establish latency and to reactivate [Whitley and Gnann, 2002]. Most of the HHVs are neurotropic and a common cause for serious acute and chronic neurological disease of the central nervous system (CNS); this can be monophasic, recurrent or chronic. Infection with each herpesvirus produces different clinical features and imaging abnormalities, and many HHV infections can now be treated [Gilden et al. 2007].
Neurological disease (encephalitis, meningitis, myelitis, cerebellitis) either occurs during primary infection or during the course of virus reinfection or reactivation. Cerebrospinal fluid (CSF) analysis is a key tool in the diagnosis of viral CNS infection; polymerase chain reaction (PCR) analysis has revolutionized this [DeBiasi et al. 2002; Boivin, 2004], replacing diagnostic uncertainties and avoiding invasive brain biopsy [Schmutzhard, 2001]. Clinical presentation and findings on magnetic resonance imaging (MRI) usually complete the diagnosis. In cases of immunosuppression [e.g. corticosteroids, chemotherapy, transplantation, human immunodeficiency virus (HIV) infection] the clinical manifestations may be atypical and the diagnosis therefore especially challenging. In patients with suspected viral encephalitis treatment with acyclovir should be started empirically and only discontinued when HSV is ruled out [Roos and Tyler, 2010].
Herpes simplex virus type 1
On the basis of DNA homologies, serological typification and clinical symptoms, one can differentiate two serotypes: HSV-1 and HSV-2. HSV encephalitis (HSVE) is the commonest fatal sporadic encephalitis in humans [Whitley, 1990]. In immunocompetent adolescents and adults the disease is caused in about 90% of cases by HSV-1, leading to focal encephalitis. HSV-2 is associated with HSVE in immunocompromised patients, in whom infection often progresses to a disseminated extent [Kennedy and Chaudhuri, 2002]. Furthermore, HSVE does not occur more frequently in immunocompromised patients. Untreated, HSVE has an extremely high mortality rate at about 70%, which decreases in treated patients to 19%, but more than 50% survive with moderate or severe neuropsychiatric sequelae [Jakob et al. 2008]. The neuropathological picture of HSVE is characteristic, consisting of acute necrotizing encephalitis that almost always localizes, often asymmetrically, to the orbitofrontal and temporal lobes with involvement of the cingulate and insular cortex [Kennedy and Chaudhury, 2002]. The mechanism by which HSV-1 infects the CNS to cause encephalitis has not been definitively established. One possibility is that the virus spreads from the olfactory mucosa through the crib-r-iform plate of the ethmoid bone into the anterior fossa. Latent virus in the trigeminal ganglia might also reactivate and spread via tentorial nerves, which innervate the meninges of the anterior and middle cranial fossa. Most humans harbor latent HSV-1; this has been found to be restricted to the cranial nerves [Gilden et al. 2007].
Unfortunately, the clinical picture of HSVE is nonspecific; the disease may follow a brief influenza-like illness or begin abruptly. The most prominent symptoms are fever, headache, confusion, altered consciousness, memory loss, or personality changes. Highly suspicious for HSVE are prodromal symptoms of an upper respiratory tract infection, in combination with neurological findings, which are related to dysfunction of the frontotemporal lobes.
Examination of the CSF is indicated, however the findings are nondiagnostic, being similar in patients with confirmed disease or diseases that mimic HSVE. The characteristic CSF profile consists of lymphocytic pleocytosis, a mildly elevated protein level and a normal glucose level, but pleocytosis might also be absent. Approximately 5–10% of patients have a normal CSF formula on first evaluation [Whitley, 2004], however if neurological symptoms remain despite an unsuspicious initial CSF, lumbar puncture should be repeated. Recent studies by Jakob and colleagues (oral communication) found that, although CSF analysis revealed a normal cell count, HSV DNA was detected in all samples by PCR. HSV-PCR should be used as the method of choice to detect HSV genomes as early as possible. With a sensitivity of 95% and a specificity of 100% [Baringer, 2008] the PCR method has become a diagnostic gold standard in HSVE. PCR has replaced the detection of intrathecal production of anti-HSV antibodies, which was the main noninvasive diagnostic tool for many years. The optimum time span for obtaining a positive HSV-PCR is probably between 2 and 10 days after onset of the illness [Davis and Tyler, 2005]. After 7 days of treatment with acyclovir, HSV genomes are rarely detectable in the CSF [Kessler et al. 1994].
When PCR results are pending, noninvasive neurodiagnostic studies support a presumptive diagnosis of HSVE, including electroencephalography (EEG) and imaging. Focal changes in the EEG are characterized by spike and slow-wave activity, as well as periodic lateralized epileptiform discharges arising from the temporal lobe [Upton and Grumpert, 1970]. These changes usually involve one side at the beginning of the disease and then spread to the contralateral temporal lobe as the disease evolves. The sensitivity of the EEG is approximately 84%, but the specificity is only 32.5% [Whitley, 2004]. Computed tomography (CT) scanning is usually normal within the first 4–6 days, MRI being much more sensitive, demonstrating high-signal intensity lesions on T2-weighted, diffusion-weighted and fluid attenuated inversion recovery (FLAIR) images earlier in the course. Brain biopsy is now rarely performed, but is indicated in unresponsive patients and in cases with profound diagnostic doubt [Steiner, 2011].
In immunocompetent patients the infection usually remains localized, as activated macrophages and lymphozytes (tumor necrosis γ production), as well as natural killer cells prevent dissemination. Here the T-cell-mediated defense seems to be more important than the humoral immune response. The immune response to HSV infection does not protect against reactivation of the latent virus, and a previous HSV-1 infection does not protect against one induced by HSV-2 (Jakob and Handermann, 2012).
In immunocompromised patients HSVE has been documented in single cases [Li and Sax, 2009] and therefore cannot be regarded as opportunistic infection. The case reports suggest that unusual clinical and neuropathological features characterize the disorder in this population. The infection may be atypical or more aggressive in patients with immunosuppression than in immunocompetent patients [Price et al. 1973]. A serious disease course mainly occurs in patients with a deficiency in cell-mediated immunity such as T-cell deficiency, corticosteroid therapy and HIV infection [Grover et al. 2004]. Delayed diagnosis and therapy may result in poor outcome or death [Kimberlin, 2007]. The incidence of HSVE in immunocompromised patients may be underestimated. Schiff and Rosenblum reported proven HSVE in two patients with cancer and in two HIV-infected individuals [Schiff and Rosenblum, 1998]. HSVE was not suspected in any of these cases before death and the CSF showed no pleocytosis.
Most cases of childhood HSVE remain unexplained. However, Casanova and colleagues elucidated a genetic etiology. They hypothesized that HSVE is a genetically heterogeneous disease, involving a collection of single-gene inborn errors of immunity to HSV-1 in the CNS during the course of primary infection. Deficiencies of UNC-93B [Casrouge et al. 2006], TLR393B [Guo et al. 2011] and TRAF3 [Pérez de Diego et al. 2010] have been found in children with HSVE who were otherwise healthy.
The first antiviral drug reported as efficacious therapy of HSVE was idoxuridine, which was soon proven both ineffective and toxic [Chien et al. 1975]. Subsequent therapeutic trials defined vidarabine as a useful medication for the management of biopsy-proven HSVE [Whitley et al. 1981]. However, the current therapy of choice is acyclovir and should be initiated at the faintest suspicion of HSVE. The drug is administered at a dosage of 10mg/kg every 8 h for 14–21 days. Acyclovir resistance has not been shown to occur in immunocompetent patients [Griffiths, 2001] but has been reported in the immunocompromised [Christophers et al. 1998]. Whether intravenous acyclovir treatment should be followed by the use of valacyclovir orally for an extended period is an issue that has yet to be resolved. In immunocompromised patients it is important to exclude CMV as a cause of encephalitis, as this should be treated with ganciclovir [Balfour, 1999].
The use of corticosteroids as an adjunct treatment for HSVE is controversial. However, when encephalitis is complicated by severe cerebral edema with MRI evidence of midline shift, high-dose steroids (dexamethasone) might have a therapeutic role [Musallam et al. 2007; Meyding-Lamadé et al. 2003].
The utility of adjunctive corticosteroid therapy in HSVE is about to be evaluated in a multicenter, multinational, randomized, double-blind, placebo-controlled trial (GACHE trial) [Martinez-Torres et al. 2008].
Although acyclovir has reduced mortality from HSVE, most survivors have persistent neurological symptoms. Despite early diagnosis, treatment and a promising early beneficial, outcome cognitive impairment remains the main problem.
Herpes simplex virus type 2
Neonatal HSVE results when HSV is transmitted from the mother to her offspring during childbirth, the majority of which are primary HSV-2 infections. HSV-2 accounts for about 80% of cases in the newborn, but after the neonatal period HSV-2 is an uncommon cause of encephalitis, being responsible for less than 10% of cases [Tang et al. 2003]. In adults HSV-2 usually causes uncomplicated genital herpes, but occasional cases of neurological involvement are identified, ranging from meningitis, which may be recurrent, to radiculomyelitis and rarely encephalitis [Dennett et al. 1997]. CNS disease alone occurs in one-third of all infants with neonatal HSV infection. In cases of neonatal HSVE due to HSV-2 exposure during delivery, a latency of typically 11–17 days following delivery has been described prior to clinical presentation [Whitley et al. 1980].
Clinical manifestations of HSV encephalitis include seizures, fever, lethargy, irritability, tremors, poor feeding, temperature instability, bulging fontanelle and pyramidal tract signs [Arvin et al. 2006]. Neurodiagnostic studies of value in CNS HSV infection include CSF examination, EEG and brain-imaging techniques. PCR is the gold standard; for limitations see the previous section. Diffusion-weighted MRI seems to be a more sensitive tool for demonstrating early CNS lesions in neonatal HSVE than conventional MRI [Kubota et al. 2006; Vossough et al. 2008].
HSV encephalitis in the newborn typically involves the cerebral cortex, not with a focal pattern but widespread and diffuse. One case of predominant brainstem and cerebellar involvement secondary to HSV-2 infection in a neonate has been reported [Pelligra et al. 2007].
With current antiviral therapy (high-dose acyclovir, 20 mg/kg every 8 h for 21 days), the mortality rate from neonatal CNS HSV disease is 4%; however, 69% of survivors are left with neurological sequelae [Kimberlin, 2003]. HSV-1 is associated with a better overall prognosis than HSV-2 as the cause of neonatal encephalitis [Toth et al. 2003]. Prophylactic acyclovir is under clinical investigation to improve the outcome of those who survive neonatal HSV disease.
Human immunodeficiency virus
HSV-2 seroprevalence among HIV-1 infected people ranges from 70% to over 90% [Weiss, 2004]. HSV-2 reactivation is common and often asymptomatic among people with HIV-1 infection [Posavad et al. 2004]. Plasma and genital HIV-1 levels increase during HSV-2 reactivation [Schacker et al. 2002], suggesting that HSV-2 reactivation enhances HIV-1 replication. Possible mechanisms are binding of HSV proteins to the HIV-1 long terminal repeat, elevation of proinflammatory cytokines, or infiltration of HIV-1 target cells in the genital tract [Moriuchi et al. 2000; Rebbapragada et al. 2007]. Given the strong relationship between higher plasma HIV-1 concentrations and faster HIV-1 disease progression, HSV-2 suppression has been considered a potential strategy to reduce HIV-1 levels and slow HIV-1 disease progression; however, the results are conflicting. Randomized trials among people with dual HIV-1/HSV-2 infection who are not taking highly active antiretroviral therapy (HAART) found that daily HSV-2 suppressive therapy with acyclovir or valacyclovir for 8–12 weeks reduced plasma HIV-1 levels [Nagot et al. 2007; Delany et al. 2009]. Similar results were found over a follow-up period of 24 months [Jairam et al. 2010].
Another study reported that treatment with acyclovir does not significantly prolong progression to acquired immune deficiency syndrome (AIDS) among people who are HIV positive and infected with HSV-2. The study found the major benefit in survival was due to antiretroviral therapy [Suligoi et al. 2002]. Daily acyclovir therapy did not reduce the risk of transmission of HIV-1, despite a reduction in plasma HIV-1 RNA and a 73% reduction in the occurrence of genital ulcers due to HSV-2 [Celum et al. 2010]. Nonetheless, anti-HSV-2 therapy in the management of people with HIV is fundamental for preventing sexual transmission of HIV through herpetic lesions.
Varicella zoster virus
VZV causes chickenpox (varicella) as the primary infection, becomes latent in the cranial nerve and dorsal root ganglia, and may reactivate decades later to produce shingles (herpes zoster) [Gnann and Whitley, 2002]. Varicella is characterized by a vesicular rash, usually accompanied by fever and malaise [Heininger and Seward, 2006], whereas herpes zoster is characterized by pain and rash restricted to one to three dermatomes.
Increasing age is the most important risk factor for developing herpes zoster; the other well defined risk factor is altered cell-mediated immunity. In particular, lymphoproliferative malignancies and organ transplant recipients are at very high risk [Gnann and Whitley, 2004].
VZV can also cause neurological complications, very rarely with varicella (most often a varicella cerebellitis), more frequently with herpes zoster. The main complication is postherpetic neuralgia, a neuropathic pain syndrome, which persists after the cutaneous manifestations have subsided. Acute neurological complications may affect both the peripheral nervous system (cranial neuropathies, motor radiculopathies) and the CNS (meningitis, myelitis, vasculitic encephalitis) [Gregoire et al. 2006]. The same complications can also occur in VZV infection without a rash [Koskiniemi et al. 2002]; neurological complications appear more often in immunocompromised patients.
Treatment for varicella is symptomatic. In the case of herpes zoster, antiviral treatment should be commenced within 72 h after appearance of the rash. The oral treatment with valacyclovir (1 g three times per day for 7 days) seems to be superior to the oral therapy with acyclovir (800 mg five times per day for 8 days); this is due to the three- to fivefold better oral bioavailability. Duration of the zoster-associated pain (acute pain as well as postherpetic neuralgia) is shortened by valacyclovir. Superinfection should be treated with antibiotic ointment. Pain relief can be combined with amitriptylin (e.g. 25 mg twice or three times per day) [Hacke, 2010].
VZV encephalitis (VZVE) usually occurs a few days to weeks after the onset of rash, but has been reported from days to weeks before or after skin eruption. Immunocompromised patients are clearly at increased risk. Other markers of increased risk for CNS involvement include herpes zoster in a cranial nerve dermatome or the presence of cutaneous dissemination [Jemsek et al. 1983]. The clinical presentation is often an acute or subacute delirium with few focal neurological signs. Other findings can include headache, meningism, fever, ataxia and seizures, all of which are not specific for VZVE. Few studies have examined the brain changes and cognitive outcomes associated with VZVE. Existing studies show considerable variability in terms of the brain regions affected, the severity of neurological damage, and the preferably affected cognitive domains [Bangen et al. 2010].
MRI of the brain shows large and small ischemic or hemorrhagic infarcts – often both – of cortical and subcortical grey and white matter. Abnormalities in particular at the grey–white matter junctions should suggest the possibility of VZV infection [Gilden et al. 2009].
CSF examination usually reveals a mild mononuclear pleocytosis, a normal or elevated level of protein and a normal level of glucose [Gilden et al. 2000]. Moreover, the pleocytosis should be interpreted with caution, as it is also present in approximately half of the patients with uncomplicated herpes zoster [Haanpaa et al. 1998]. As in HSV the method of choice is CSF PCR for VZV DNA [DeBiasi et al. 2002]. In patients with HIV infection, CSF PCR for VZV DNA may have utility in monitoring therapeutic response and in predicting the outcome of VZV meningoencephalitis [Cinque et al. 1997].
In immunocompetent patients a booster reaction occurs and often VZV immunoglobulin M (IgM) can be detected. Cellular immunity plays a crucial part in immune control [Hengel, 2012]. Pathogenesis of VZV infection requires evasion of innate immunity and limited secretion of antiviral cytokines. Recent studies reported that its immediate–early protein ORF61 antagonizes the β-interferon pathway via direct interaction and may contribute to its pathogenesis [Zhu et al. 2011].
Since the introduction of real-time PCR in 2003, VZV is a more commonly diagnosed cause of viral CNS infection. Quantitative PCR is more sensitive in detecting VZV in the CSF than the previously used qualitative PCR method. Persson and colleagues reported that patients with encephalitis and meningitis have higher viral loads than patients with other CNS syndromes caused by VZV [Persson et al. 2009]. The majority of them have neurological sequelae when seen in follow up.
Doses of acyclovir in VZVE are similar to HSVE and treatment should be continued for 3 weeks. The use of large doses of corticosteroids as adjunct therapy in acute viral encephalitis is controversial and probably has the best evidence for VZVE, as this is understood to be caused by cerebral vasculitis [Haeusler et al. 2002; Gilden et al. 2009]. Large vessel arterial disease (granulomatous arteritis) occurs predominantly in immunocompetent patients, whereas encephalitis, mediated by small vessel disease, is virtually exclusively found in immunodeficient hosts [Gilden et al. 2000]. The duration of steroid treatment should be short (between 3 and 5 days) to minimize adverse effects; it should not exceed 1 week as long-term treatment can potentiate virus infection [Gilden et al. 2009].
In 1998 the World Health Organization recommended that routine childhood varicella vaccination should be considered in countries where the disease is a relatively important public health and socioeconomic problem, where the vaccine is affordable and where high (85–90%) and sustained vaccine coverage can be achieved [World Health Organization, 1998]. Since this vaccination policy was implemented in the United States, figures for incidence, morbidity and mortality related to VZV have decreased. Current recommendations are to routinely vaccinate children at age 12–15 months and to vaccinate all healthy individuals aged 13 years and above without evidence of anti-VZV immunity [Marin et al. 2007].
Example of magnetic resonance imaging findings in herpes simplex virus encephalitis (coronary T2-weighted fluid attenuated inversion recovery) (with acknowledgements to Professor Dr B. Kress, Head, Department Neuroradiology, Krankenhaus Nordwest, Frankfurt, Germany).
Example of magnetic resonance imaging findings in zoster neuritis (axial T1 weighted after gadolinium contrast) (with acknowledgements to Professor Dr B. Kress, Head, Department of Neuroradiology, Krankenhaus Nordwest, Frankfurt, Germany).
(a) Example of magnetic resonance imaging (MRI) finding of cerebral lymphoma (axial T2 weighted). (b) Example of MRI finding in cerebral lymphoma (axial T1 weighted after gadolinium contrast) (with acknowledgements to Professor Dr B. Kress, Head, Department of Neuroradiology, Krankenhaus Nordwest, Frankfurt, Germany).
However, recommendations for varicella vaccination in Europe vary, with the majority of countries not following the WHO recommendations for routine vaccination, despite the proven health benefits of varicella vaccination. Most European countries (except Germany and Greece) have delayed the introduction of the vaccine into their immunization schedules [Pinot de Moira and Nardone, 2005]. Considerable care needs to be taken with vaccination in immunocompromised patients since complications have been reported from vaccines given to children with undiagnosed immunodeficiencies [Leung et al. 2004].
Epstein-Barr virus
EBV causes infectious mononucleosis (IM) and can also be responsible for infections of the CNS. It is one of the most successful viruses, infecting over 90% of human beings in the first decades of life and persisting for the lifetime of the person [Gratama and Ernberg, 1995]. Primary infection usually occurs through infected saliva and is asymptomatic in young children, but in up to 40% of adolescents and adults it leads to IM. This is an acute and usually self-limiting lymphoproliferative disease, characterized by sore throat, fever, adenopathy and splenomegaly [Kutok and Wang, 2006]. CNS symptoms may occur shortly before, during or after IM, as well as following acute EBV infection in the absence of symptomatic IM. CNS complications include meningitis, encephalitis, cerebellitis, polyradiculomyelitis, transverse myelitis, cranial and peripheral neuropathies [Hoover et al. 2004].
EBV-related CNS infections may be classified into two groups: CNS syndromes associated with primary EBV infection or reactivated infection; and those associated with chronic active EBV infection [Fujimoto et al. 2002]. As EBV does not become latent in neurons or other nonlymphoid cells, a CNS infection following EBV reactivation is likely to occur at extraneural sites with subsequent spread of the virus to the CNS via infected lymphocytes [Weinberg et al. 2002].
The clinical presentation of encephalitis caused by EBV does not differ from viral encephalitis in general, meaning it is not specific. Patients can present with seizures, coma, personality changes, focal brainstem or cerebral findings, cerebellar ataxia or distortions of perception. Although EBV encephalitis may involve any area of the brain, the cerebellum is commonly involved; most patients present with abnormalities of the gait [Hoover et al. 2004]. Unusual complications include extrapyramidal involvement, acute aqueductal stenosis, syndrome of inappropriate antidiuretic hormone secretion, Reye syndrome or acute fatty degeneration of the liver, accompanied by encephalopathy.
EBV-related neurological conditions can be diagnosed by PCR detection of EBV DNA in the CSF, as well as changes in EBV serology, consistent with an acute infection [Tselis et al. 1997]. MRI is considerably sensitive and can reveal lesions, particularly on T2-weighted images, when CT findings are normal. In patients with encephalitis EEGs frequently show generalized slowing with occasional bursts of activity. Regarding treatment, acyclovir and corticosteroids are recommended for EBV CNS infections, though the efficacy of both has not been established.
The humoral response to EBV infection involves both viral specific and nonspecific antibodies. Detection of heterophil antibodies, along with symptoms of IM and atypical lymphocytes, is diagnostic of acute EBV infection. Specific antibody responses are more reliable than heterophil antibodies. Detection of IgM antibodies to virus capsid antigen, seroconversion of ‘early antigen’, as well as the absence of anti-Epstein-Barr nuclear antigen (EBNA) antibodies – which usually appear later in the course of the infection – characterize an acute EBV infection [Hoover et al. 2004; Rösen-Wolff, 2012]. As part of the innate immune response, natural killer (NK) cells have an important role in the control of primary EBV infection. They eliminate infected B cells and augmenting the antigen-specific T-cell response via release of immunomodulatory cytokines. The magnitude of the NK cell response may ultimately determine whether primary EBV infection has a clinical outcome [Williams et al. 2005].
EBV has been found to be associated with numerous cancers, including Burkitt’s lymphoma [Epstein et al. 1964], Hodgkin’s disease [Weiss et al. 1989] and nasopharyngeal carcinoma [Zur Hausen et al. 1970]. EBV-associated cancers represent another important cause of EBV-related CNS disease. These disorders are particularly common in immunocompromised hosts, including patients with AIDS and transplant recipients. Patients with AIDS have 10–20 times as many circulating EBV-infected B cells as healthy people. T cells from patients with AIDS suppress EBV-infected B cells less effectively than cells from normal controls [Birx et al. 1986].
The pathogenesis of EBV-associated neurological disease is thought to be different from that of HSV CNS disease. As viral proteins are present in the brain of patients with HSVE, brain biopsies from patients with EBV encephalitis do not show viral nucleic acids or proteins [Pedneault et al. 1992]. Therefore autoimmune mechanisms are probably responsible for the pathogenesis of EBV-associated CNS disease [Ito et al. 1994].
Recent studies have shown an association between EBV infection and multiple sclerosis (MS), suggesting a role of EBV in induction and pathogenesis of MS. There is considerable evidence that EBV infection is a strong risk factor for the development of MS and that EBV infection itself increases the risk of MS [Levin et al. 2010]. One of the important questions in this context is whether EBV-infected B lymphocytes are present within the lesions of patients with MS. Previous studies provided controversial results, as they detected EBV reactivity in B cells, either in most of the MS cases and lesions or just in rare cases [Lassmann et al. 2011].
Cytomegalovirus
Cytomegalovirus is the largest herpesvirus and causes a variety of clinical manifestations, ranging from unapparent to fatal. Infection usually occurs early in life and remains latent in immunocompetent individuals. The sites of latency are not fully understood, PCR studies of blood cells have detected CMV in monocytes/macrophages [Taylor-Wiedeman et al. 1991]. Antibodies of IgG class against CMV can be found in approximately 60% of adults in developed and virtually 100% in developing countries [Griffiths and McLaughlin, 2004]. Most primary infections are asymptomatic. However, immunocompromised hosts, particularly organ transplant recipients or those suffering from AIDS, can develop severe clinical disease, either from a primary infection or reactivation of a latent CMV infection [Van der Bij and Speich, 2001]. Before the advent of HAART, CMV infection was the most frequent opportunistic infection of the CNS in patients with AIDS. Since then, both infection and disease have decreased dramatically in countries where HAART is available [Griffiths and McLaughlin, 2004]. CMV encephalopathy can predominate as the cause of death in a child who acquires both CMV and HIV from its mother [Belec et al. 1990]. This suggests an interaction between HIV and CMV, which seems to cause a further suppression of the immune system, leading to an escalation of the disease. Results of experimental studies in an animal model support the hypothesis that a depressed immune system is necessary for CMV to invade the mature brain [Reuter et al. 2004].
In immunocompetent hosts the virus remains efficiently controlled and several components of the immune system are known to play a role. Cellular immunity holds the essential role; antibodies have a supporting function. In infected cells the CMV infection acts lytically; multiple genetic functions help the virus to partially escape immune control. CMV can keep major histocompatibility complex mediated antigen presentation under control and antagonize the T-cell response to it. CMV can also attenuate the recognition of infected cells via NK cells, as well as interrupt numerous intracellular signal cascades and with these the effect of interferons and antiviral cytokines. The viral immune evasion can be compensated by specific capacities of the immune system, but this requires an intact cellular immunity. Activated T lymphocytes, CMV-specific IgM and IgA characterize an acute infection [Hengel, 2012].
Transmission of infection from one individual to another requires direct contact. Sources of virus include saliva, blood, vaginal secretions, semen and breast milk. Further, CMV can be transmitted iatrogenically by organ allografts or blood products. If a CMV IgG-seropositive patient becomes immunosuppressed, reactivation of CMV is common and can be found in 50% of transplant recipients at some time after transplantation. The donor organ can also transmit CMV to seronegative individuals. Preexisting immunity provides moderate protection against CMV disease. Current guidelines recommend 3–6 months of prophylaxis with valganciclovir for high-risk patients [Humar and Snydman, 2009].
In congenital CMV infection the fetus is infected transplacentally; perinatal infection occurs via maternal genital secretions or breast milk. Approximately 7% of cases have symptoms at birth of cytomegalic inclusion disease (microcephaly, hepatosplenomegaly, thrombocytopenic purpura, hearing loss, intracranial calcifications); their prognosis is very poor [Stagno, 2000]. Sensorineural hearing loss affects 40–60% of symptomatic infants and 7–15% of asymptomatic infants [Williamson et al. 1992]. A randomized, controlled trial has reported a significant decrease in loss of hearing when neonates with congenital CMV infection were treated with ganciclovir [Kimberlin et al. 2003].
Neurological complications include encephalitis, myelitis and polyradiculopathy. In the case of encephalitis no unique clinical features are produced, and also typical laboratory or neuroimaging findings are absent. This makes the diagnosis of CMV encephalitis (CMVE) very difficult. Moreover, it is clinically indistinguishable from HIV dementia. Nonspecific signs of encephalopathy (e.g. confusion, lethargy, disorientation, perhaps seizures) represent the main clinical feature of the infection, therefore in some cases diagnosis requires autopsy [Griffiths and McLaughlin, 2004]. Pathological studies suggest that the most common sites of CMV infection of the brain are the basal ganglia, diencephalon and brainstem. Furthermore, the virus is thought to be able to induce heterogeneous structural abnormalities with variable degrees of tissue response. This could explain its localization in the nervous system without significant clinical sequelae [Vinters et al. 1989]. Therefore CMVE should be considered in the differential diagnosis of encephalitis in immunosuppressed patients; diagnosis is best made by PCR. A PCR search for the CMV genome in the CSF in these patients is suggested, even if slight clinical symptoms are present. Ganciclovir and valganciclovir provide effective preventative (after solid organ transplantation) and treatment options, targeting the viral DNA polymerase. Treatment-related neutropenia and the emergence of ganciclovir-resistant virus are the limitations to current management strategies. Ganciclovir-resistant CMV is most commonly seen in CMV-negative recipients of CMV-seropositive lung transplants [Boivin et al. 2005]. Foscarnet and cidofovir are the two most common alternatives, but both are nephrotoxic. AIC246 belongs to the novel chemical class of 3, 4 dihydroquinazolinyl-acetic acids and is a promising new drug candidate, currently in phase IIb development [Lischka et al. 2010]. The first successful use of AIC246 in the treatment of CMV disease in a lung transplant recipient with disseminated multidrug resistance has been reported [Kaul et al. 2011].
The goal of CMV vaccine is to prevent congenital infection and its sequelae. Development of a vaccine was listed as a top priority in the United States by a committee of the Institute of Medicine in 2001 [Stratton et al. 2001]. Although the first clinical trials of a CMV vaccine took place more than 30 years ago, an effective vaccine for the prevention of CMV infection remains elusive. In 2009, a phase II clinical trial with CMV glycoprotein B vaccine (plus MF59 adjuvant) showed a reduction of infection in women of child-bearing age as well as congenital infection. A limitation of these results is certainly the very low sample size [Pass et al. 2009]. Further studies will have to demonstrate reasonable safety and efficacy before this vaccine can be recommended in preventing CMV infection in young women and congenital CMV in their infants.
Table 1 summarizes the data on the HHVs.
Summary of human herpesviruses.
In immunocompromised patients.
ASI, antibody specific index; CMV, cytomegalovirus; EBV, Epstein-Barr virus; HHV-6, human herpes virus 6; HSV, herpes simplex virus; ME, meningoencephalitis; PCR, polymerase chain reaction; VZV, varicella zoster virus.
Human herpesvirus 6
This virus was first discovered in 1986 in patients with AIDS, therefore initial investigations focused on its role as a possible pathogen in people with HIV infection. On the basis of an observed tropism for B cells, the virus was named human B-lymphotropic herpesvirus (HBLV) [Salahuddin et al. 1986]. Ongoing investigations indicated that the virus was most likely to be found and to grow in CD4 lymphocytes [Takahashi et al. 1989]. The name was changed to HHV-6, a name independent of the cell tropism and in accordance with guidelines established for the taxonomy of viruses.
In 1988, HHV-6 was identified as the cause of the common child disease exanthema subitum (roseola infantum or sixth disease). This disease generally occurs within the first 2 years of life, is characterized by intense fever for 2–3 days, followed for a minority of infants (25–30%) by a nonvesicular cutaneous rash on the trunk and back [Yamanishi et al. 1988]. HHV-6 is one of the most widespread herpesviruses, infecting the human population with a prevalence approaching 100%. The most probable route for transmission is through saliva, either from mother to child or between children [Mukai et al. 1994]. Two types (A and B) can be identified; no diseases have clearly been linked to HHV-6A infection, whereas HHV-6B is responsible for exanthema subitum. Complications of this primary infection include seizures, hemiplegia, meningoencephalitis or residual encephalopathy, illustrating HHV-6 neurotropism. As with all the herpesviruses, HHV-6 establishes latency in monocytes or macrophages [Kondo et al. 2002], as well as in brain tissue [Chan et al. 2001]. Reactivation occurs preferably in immunocompromised people but also in immunocompetent hosts. HHV-6 is known to reactivate frequently during acute infections with other viruses, especially with other herpesviruses [Braun et al. 1997]. Cases of HHV-6-associated encephalitis in immunocompetent patients have been reported [Denes et al. 2004; Isaacson et al. 2005], even one case of relapsing HHV-6 encephatis [Sawada et al. 2007]; however, their pathogenesis remains unclear. An undeniable possibility is that the patients had an immunocompromising disease that had not been diagnosed.
The lifelong consequences of infection with HHV-6 are still not well known. HHV-6 DNA can be detected at several body sites years after primary infection; it seems to have a strong affinity to the CNS and shifts there during primary infection at a high rate [Caserta et al. 2001]. The clinical relevance of reactivation is not completely understood. However, in transplant recipients HHV-6 is recognized as an opportunistic pathogen.
In adults the virus has been associated with a broad range of neurological disorders, including multiple sclerosis [Cirone et al. 2002], temporal lobe epilepsy [Donati et al. 2003] and post-transplant acute limbic encephalitis (PALE). Consistent PALE clinical features include marked anterograde amnesia, seizures or temporal lobe EEG abnormalities and mild CSF pleocytosis. The syndrome is accompanied by MRI abnormalities such as bilateral, nonenhancing, medial temporal lobe T2/FLAIR/diffusion-weighted imaging hyperintensities, sharply demarcated by parahippocampal gyrus sparing [Wainwright et al. 2001; MacLean and Douen, 2002], and is associated with a short interval from transplantation (15–60 days) [Seeley et al. 2007]. PALE could reflect HHV-6 reactivation as the virus resides latently in up to 85% of non-neurological autopsy control brains [Chan et al. 2001] and has a predilection for the medial temporal lobes [Donati et al. 2003]. Differential diagnosis on MRI may include HSVE, other viral encephalopathies, CNS invasion of lymphoproliferative diseases, paraneoplastic or drug-induced encephalopathy. However, the combination of characteristic mesial temporal lobe involvement on brain imaging plus the immunocompromised condition (especially from 2 to 4 weeks after transplantation) under the preventive administration of acyclovir is highly suggestive of HHV-6-associated encephalopathy.
Immune responses to HHV-6 include regulation of cytokine production, downregulation of T-cell receptors and impairment of T-cell activation, as well as inhibition of dendritic cell maturation and functions [Sullivan and Coscoy, 2008; Smith et al. 2005]. HHV-6 can interfere with the function of the host immune system through a variety of mechanisms. The primary target for HHV-6 replication is the CD4+ lymphocyte. HHV-6A, but not B, also replicates in various cytotoxic effector cells, such as CD8+ T cells, γδ T cells and NK cells. HHV-6 induces dramatic functional abnormalities, including a selective suppression of interleukin 12, which seems to be mediated by the engagement of the primary HHV-6 receptor, CD46. By modulating specific antiviral immune responses, HHV-6 can facilitate its own spread and persistence in vivo, as well as enhance the pathogenic effects of other agents, such as HIV [Lusso, 2006].
As with all the herpesviruses the main diagnostic tool is HHV-6 DNA PCR assay on the CSF, its detection assuming virus replication in the CNS. This has been questioned in view of the phenomenon of HHV-6 chromosomal integration [Clark et al. 2006] because this is the only HHV found integrated into host chromosomes [Tanaka-Taya et al. 2004]. Thus, when assessing the significance of viral DNA in CSF and its relationship to neuropathogenesis in the immunocompetent patient, chromosomal HHV-6 integration should be distinguished from primary infection and virus reactivation [Ward et al. 2007].
Regarding treatment, HHV-6 infections in immunocompetent children are self limiting and do not require treatment. In immunocompromised individuals, however, reactivation of latent virus may cause life-threatening complications. As an effective vaccine is not available, safe and reliable drugs are needed. To date, standardized therapy of HHV-6 infection in immunosuppressed patients is lacking and no drug has been widely approved for treatment of HHV-6-associated disease.
Because HHV-6 lacks a thymidine kinase, it is insensitive to acyclovir at achievable serum concentrations, making the use of acyclovir inadequate [De Clercq et al. 2001]. Ganciclovir is currently used as first-line therapy for HHV-6 encephalitis. Foscarnet and ganciclovir demonstrated the most antiviral activity, but only foscarnet and cidofovir were effective at inhibiting HHV-6 replication in glia cells [Akhyani et al. 2006]. A synergy effect of ganciclovir plus foscarnet has been reported [Troy et al. 2008]. Many patients under immunosuppressive therapy already receive acyclovir or ganciclovir prophylaxis to prevent HSV and CMV reactivations respectively. For ganciclovir, this prophylaxis also seems effective in preventing HHV-6 reactivation [Rapaport et al. 2002].
Human herpesvirus 8
HHV-8 is the eighth HHV to be identified and has been classified as a member of the γ-herpesvirus subfamily [Moore et al. 1996]. Although it does not cause CNS disease, it should be mentioned because of the strong relation to immunosuppression.
In 1994 Chang and colleagues identified herpesvirus-like DNA fragments in a Karposi’s sarcoma (KS) skin lesion from a patient with AIDS, which was then called KS-associated herpesvirus (KSHV or HHV-8) [Chang et al. 1994]. Subsequently, other investigators described four clinical variants of Kaposi’s sarcoma, which had identical histological features, but developed in specific populations and had different sites of involvement and rates of progression: classic, AIDS associated, post transplantational (iatrogenic or immunodeficient) and African (endemic) subtypes [Antman and Chang, 2000]. KS occurs in the skin, oral cavity, gastrointestinal tract, lung, liver, lymph node, etc. Skin lesions of KS are most common. It is clear, that the presence of KSHV is the primary and necessary factor in the development of this tumor. In addition, immunosuppression in the host appears to be an important cofactor. KSHV is clearly associated with KS, body-cavity-related B-cell lymphoma (primary effusion lymphoma), and some plasma-cell forms of multicentric Castleman’s disease.
KSHV is distributed all over the world, and there are many individuals with KSHV infections. Therefore, a low KSHV titer, as detected by PCR, does not mean that a disease is associated with KSHV infection. Because latency-associated nuclear antigen-1 (LANA-1) is always expressed in KSHV-infected cells, LANA-1 immunohistochemistry is a powerful and confirmative tool to detect KSHV-infected cells in pathological samples [Fukumoto et al. 2011].
Most primary KSHV infections appear to be asymptomatic. Clinical and epidemiologic studies have shown that, in healthy adults, there is immunological control of KSHV infection. In HIV-seropositive patients, the incubation period for diseases caused by KSHV infection largely depends on the host’s immune status rather than on the duration of KSHV infection [Osman et al. 1999]. Underlining the importance of the immune status of the host is the finding that both KS and body-cavity-related B-cell lymphoma have dramatically responded to HAART in patients with AIDS [Winceslaus, 1998].
Treatment modalities comprise local therapy, for example, surgery, radiotherapy and local chemotherapy. Patients with widespread disease may need systemic chemotherapeutic or immunological medication. Positive results have been found for pegylated liposomal doxorubicin, danaurubicin, paclitaxel and interferon α [Di et al. 2008; Brambilla et al. 2008]. In patients with iatrogenic KS, immunosuppressive medication may be reduced or modified with the possibility of grafts being rejected with insufficient immunosuppression [Montagnino et al. 1994].
Conclusion
All of the discussed HHVs can cause serious neurological disease of the CNS through primary infection or following virus reactivation, especially in an immunocompromised host.
Recognition of the clinical patterns and imaging characteristics of disease produced by different herpesviruses is important because infection by many of the herpesviruses can be treated successfully. Early diagnosis and proper treatment are essential for a favorable outcome. Dealing with herpesvirus infections in immunosuppressed patients is particularly challenging. The new immunosuppressive approach to neuroinflammatory disease has enabled a remarkable new generation of powerful therapies but will increase the complexity of possible risks to patient safety. Systematic structured risk management will therefore be needed as some of the key adverse effects of immunosuppression can be prevented by careful patient selection and preemptive vaccination or treatment.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare no conflicts of interest in preparing this article.