Progressive multifocal leukoencephalopathy: an unexpected complication of modern therapeutic monoclonal antibody therapies

Abstract

Progressive multifocal leukoencephalopathy (PML) is a rare demyelinating disorder of the central nervous system, caused by the reactivation of the ubiquitous JC virus. PML usually occurs during severe immunosuppression, and the most common causes are represented by human immunodeficiency virus infection, lymphoproliferative disorders and other forms of cancer. Recently, the introduction of monoclonal antibodies (e.g. natalizumab, rituximab, efalizumab) in the treatment of several dysimmune diseases such as multiple sclerosis, rheumatoid arthritis, psoriasis and systemic lupus erythematosus, has led to an increased incidence of PML. This phenomenon has had severe consequences, leading, for example, to the withdrawal from the market of Efalizumab, and important restrictions in the use of the other compounds, all of which are characterized by high efficacy in improving prognosis and quality of life. In this review we will discuss clinical, laboratory and imaging findings of PML. In addition, proposed pathogenetic mechanisms promoting the reactivation of JC virus in the context of treatment with monoclonal antibodies will be described.

Introduction

The definition ‘Progressive Multifocal Leukoencephalopathy’ (PML) was first used in 1958 to describe a fatal demyelinating central nervous system (CNS) disease in patients with lymphoproliferative disorders [1]. In 1971, the virus responsible for the disease was isolated and named JC virus (JCV) after the initials of the patient from whom the virus was first isolated [2]. After 40 years, several aspects of the disease are well known but many questions regarding the pathogenesis remain unanswered and no effective treatment is available. In the pre-human immunodeficiency virus era PML was anecdotally reported in some patients with myelolymphoproliferative disorders or other forms of cancer, but its incidence significantly increased with the advent of acquired immunodeficiency syndrome, affecting 0.7/1000 persons/year [3]. After the introduction of antiretroviral treatment the incidence (0.07/1000 persons/year) [3] and the mortality of the disease reduced significantly [4,5]. Recently, several cases of PML related to the use of new immunomodulatory compounds for the treatment of several dysimmune diseases were reported, leading to an urgent need for the identification of risk factors and disease biomarkers.

Pathogenesis

JC virus is a double-stranded, circular DNA virus and a member of the Polyomaviridae family. The viral genome encodes six viral proteins, including two early regulatory regions (small and large T antigens), another regulatory protein acting as a viroporin (agnoprotein) and the capsid proteins (VP1, VP2 and VP3) (Fig. 1). The presence of rearrangements in the highly variable non-coding control region characterizes the so-called Mad-1 type strain, neurotropic and pathogenic, which probably derives from the non-pathogenic archetype, detected in kidneys and tonsils.

FIG. 1.

JC virus circular DNA encodes for early transcripts, including small and large T antigen and for late proteins (capsid proteins VP1, VP2, VP3 and agnoprotein).

There are two main hypotheses regarding the route of primary infection, involving either upper respiratory tract through inhalation or the gastrointestinal tract through ingestion of contaminated food and water. The finding of JCV in tonsils seems to support the former hypothesis [6], but the virus has also been detected in epithelial cells from the gastrointestinal tract [7,8] and in the oesophagus [9]. However, as JCV can infect circulating B lymphocytes, both tonsils and gastrointestinal tract may represent a site of latency rather than the entry route of the virus. After the primary asymptomatic infection, which is thought to occur in childhood, the virus remains latent in different sites, including kidney—as demonstrated by the shedding of virus in the urine of 30% of immunocompetent subjects [10]—bone marrow [11], B lymphocytes [12] and tonsils [6]. There is recent evidence that the virus enters the brain from the early phases of infection and establishes a non-productive or low-chronic infection of glial cells [13,14]. Under particular conditions, usually associated with severe immunosuppression, JCV can actively replicate into the brain, leading to PML. The site and modality of JCV reactivation are still poorly understood but the most likely hypothesis is that the virus reactivates some-where in the periphery and crosses the blood–brain barrier through circulating infected cells, such as B lymphocytes, entering the CNS where it infects astrocytes and promotes lysis of oligodendrocytes with a consequent massive demyelination, predominantly involving white matter.

Clinical Findings

The classic form of PML has a fulminating evolution usually leading to death within a few months. The onset is usually multisymptomatic and the most common clinical presentation includes motor deficits, altered consciousness, gait ataxia and visual symptoms. Atypical presentations, usually reported in human immunodeficiency virus-infected patients after the introduction of combination antiretroviral therapy, include pure cerebellar syndrome, reflecting a productive infection of granule cell neurons [15], meningitis [16], meningoencephalitis [17, 18], progressive myoclonic ataxia [19] and muscle wasting associated with extrapyramidal signs [20].

Histopathological and Imaging Findings

Histopathology findings are typically represented by demyelination, enlarged oligodendrocytes productively infected by the virus as demonstrated by appropriate staining, and bizarre astrocytes [21]. Foamy macrophages may be found at the advancing edge of the lesions, actively involved in the removal of myelin breakdown products [22]. Inflammation is usually little or absent. Conventional magnetic resonance imaging shows an extensive involvement of white matter, with multifocal, bilateral, asymmetrical lesions that appear hyperintense in T2-weighted sequences and hypointense in T1-weighted sequences (Fig. 2). The pathogenetic process starts from the subcortical white matter and evolves towards the deep peri-ventricular regions, involving also thalamus and basal ganglia, with coalescence of lesions in the mid-to-late stage [23]. Pari-eto-occipital lobes are preferentially involved [24]. Infratentorial white matter is frequently affected, with lesions located in the middle cerebellar peduncles and adjacent regions, e.g. pons and mid-cerebellum [25]. As inflammation is uncommon, oedema and contrast enhancement are atypical findings.

FIG. 2.

Axial FLAIR sequence of the brain showing hyperintense, multifocal, bilateral asymmetrical lesions of the white matter, mainly involving the posterior regions (reproduced with the courtesy of Prof. S. Bastianello).

Laboratory Findings

The diagnostic criteria recently proposed require the detection of the virus in cerebrospinal fluid by PCR [22]. The sensitivity of this laboratory technique was 72–92% in the era before combination antiretroviral therapy, with a specificity of 92–100% but this fell markedly after the introduction of antiretroviral drugs [26]. It is important, therefore, to underline that even when JCV is undetectable in cerebrospinal fluid, the diagnosis of PML should not be excluded in cases with clinical and magnetic resonance imaging findings strongly compatible with the disease, and lumbar puncture should be repeated over time and brain biopsy should eventually be taken into consideration.

PML and Immunomodulatory Treatments

In 2005 the first three cases of PML in patients treated with natalizumab were reported (two with multiple sclerosis and one with Crohn’s disease) [27–29]. Since then, PML has been described as an unexpected complication of different novel immunomodulatory therapies used in dysimmune diseases, such as natalizumab, efalizumab, rituximab and alemtuzumab. We here review epidemiological and clinical data on the occurrence of PML in patients treated with the first three compounds, and hypotheses on the related pathogenetic mechanisms.

Natalizumab is a humanized monoclonal antibody directed against the cellular adhesion molecule α4-integrin, which facilitates the entry of inflammatory cells into the CNS [30]. This mechanism has a beneficial role in multiple sclerosis reducing the inflammatory infiltrates responsible for lesion formation, but on the other hand it impedes immune surveillance across the blood–brain barrier, with a reduction of the CD4⁺ : CD8⁺ T-cell ratio within the CNS [31] that might facilitate the occurrence of PML, especially considering that glial cells are believed to represent a site of viral latency.

Other hypotheses on the pathogenesis of PML in patients treated with natalizumab include the mobilization of infected B cells and peripheral haematopoietic progenitor cells. CD34⁺ facilitated by the blockage of α4-integrin plays an important role in sequestering these cells in lymph nodes and bone marrow. Recently, though, Warnke et al. [32] did not find JCV in CD34⁺ progenitor cells or peripheral mono-nuclear blood cells of 67 patients with multiple sclerosis treated with natalizumab, weakening the aforementioned hypothesis.

After the first PML cases natalizumab was briefly with-drawn from the market for safety concerns but was rapidly brought back on the market for its evident beneficial effects on multiple sclerosis evolution. Since its reintroduction, 102 cases of PML have been reported from Biogen idec as of 4 March 2011 (55 in the USA, 42 in the European Union and five in other areas) (data on file; Biogen Idec, https://medinfo.elan.com/pdfs/220.pdf), with an overall incidence of 1.23/1000 patients. The incidence is not stable over time but it depends on the treatment duration, with a rate of 1.87 cases/1000 patients on the drug for a year or longer, rising to 2.41/1000 for those on the drug for 2 years or longer, and dropping to 1.4 for those on the drug for more than 3 years. At present, the beneficial effects of the drug are considered significantly greater than the risk of PML, but multiple sclerosis patients need to satisfy strict eligibility criteria and are carefully monitored for the duration of the treatment.

So far, the search for risk factors or early biomarkers of disease has not been conclusive: viruria, viraemia and levels of anti-JCV antibodies have given conflicting results, probably because of the use of different assays and the fact that JCV is widely diffused, regularly shed in urine and detected in the blood of immunocompetent subjects. Although some authors [33, 34] did not detect any increase in plasma viraemia in patients treated with natalizumab, others [35] reported higher JCV infection in peripheral mononuclear blood cells compared with plasma.

The research in this area is active and some clinical trials are ongoing, assessing new and more sensitive assays to detect any change in JCV replication and immune system activation and seeking reliable biomarkers.

Efalizumab, analogous to natalizumab, inhibits T-cell adhesion and diapedesis from the circulation [36]. It is a recombinant humanized monoclonal IgG1 antibody directed against the CD11a subunit of leukocyte function-associated antigen-1 expressed on T cells, B cells and monocytes, approved by the Food and Drug Administration in 2003 for the treatment of moderate-to-severe chronic plaque psoriasis. In 2009 the drug was withdrawn from the market because of the occurrence of three cases of PML out of more than 46 000 treated patients [37].

The occurrence of PML in relationship with compounds that limit the access of T lymphocytes to the CNS underlines the importance of the cellular component of the immune system in the control of JCV replication within the brain.

Interesting information on the pathogenesis of the disease indirectly comes from the analysis of PML associated with the use of rituximab, a chimeric IgG1 that links to CD20 inducing a severe depletion of B lymphocytes. This monoclonal antibody was approved in the USA in 1997 for the treatment of non-Hodgkin’s lymphoma and is now marketed also for rheumatoid arthritis, chronic lymphatic leukaemia and as an off-label treatment in other diseases such as systemic lupus erythematosus. At present, more than 70 cases of PML have been associated with the use of rituximab [38–41], predominantly in patients with lymphoproliferative disorders. The risk quantification of PML related to the use of rituximab is difficult because PML has been described in systemic lupus erythematosus, rheumatoid arthritis, lymphoma and leukaemia independently from the use of any treatment.

The occurrence of PML in the context of B-lymphocyte depletion suggests not only that B cells are not the principal vehicle for JCV to enter the brain, but also that humoral immunity might play a role in the control of JCV replication, as suggested by the description of PML in patients with congenital deficiency of humoral immunity [42]. On the other hand, B-cell depletion may impact on T-cell activity, which is known to be important in the control of JCV replication, and may stimulate the proliferation and redistribution of infected pre-B lymphocytes from the bone marrow to the bloodstream, favouring viral entry into the CNS.

Although the use of these drugs has raised significant safety concerns, rituximab and natalizumab have been shown to be very effective and drastically change the disease course. Hence, the American and European regulatory agencies, after a careful re-evaluation of the benefit : risk ratio, decided to not withdraw the drugs from the market and, instead, to define strict eligibility criteria and put effort into creating management guidelines in the case of PML occurrence. These guidelines, together with the identification of risk factors for JCV reactivation and early biomarkers, should help to identify the eligible patients with a consequent reduction of PML occurrence and improvement in the management of the disease.