Treatment-related progressive multifocal leukoencephalopathy: current understanding and future steps

Abstract

Progressive multifocal leukoencephalopathy (PML) is a rare demyelinating disorder of the brain caused by a ubiquitous polyomavirus, JC virus. PML is almost always associated with some underlying immunosuppression and acquired immune deficiency syndrome has been the most common predisposing disorder. Recently, different pharmacological agents have been demonstrated to increase the risk of PML. Therapies that predispose people to PML can be classified into three categories: therapies that uniquely increase the risk for the disorder, such as the monoclonal antibodies natalizumab and efalizumab; therapies that appear to increase the risk in individuals already at risk of PML due to pre-existing conditions, such as rituximab and mycophenolate mofetil; and therapies with a mechanism of action that might suggest a potential for increased PML risk and/or with which rare cases of PML have been observed. Unlike the latter two classes, therapeutic agents uniquely increasing the risk of PML are associated with a much greater prevalence of the disorder and a latent interval from the time of drug initiation to the development of PML. PML development with pharmacological agents has provided new insight into the pathogenesis of this devastating disorder. This review focuses on the risks of PML with multiple pharmacological agents, the proposed pathogenesis with these agents, and potential risk mitigation strategies.

Background

Progressive multifocal leukoencephalopathy (PML) was first described in 1958 by Aström and colleagues [Astrom et al. 1958]. They reported unexplained progressive white matter disorder in three patients who had underlying lymphoproliferative diseases. A review of literature by the authors revealed occasional reports of similar demyelinating disorder dating back to 1930. In 1959, a viral etiology was suggested by Cavanaugh and colleagues based on observation of inclusion bodies in oligodendroglial nuclei within PML brain tissue [Cavanaugh et al. 1959]. ZuRhein analyzed these inclusion bodies in detail by electron microscopy and suggested papovavirus as the likely viral agent [Zu Rhein, 1965]. This was confirmed by Padgett and colleagues at the University of Wisconsin [Padgett et al. 1971]. The virus was labeled JC virus (JCV) after the initials of person from whom it was first isolated (John Cunningham).

From 1958 to 1984, Brooks and Walker identified 230 cases from their own experience and from the extant English language publications [Brooks and Walker, 1984]. Ninety-five percent of patients in this series had a recognized underlying condition that predisposed them to PML. Hematological malignancies, predominantly B-cell disorders, accounted for most of the underlying disorders. Other hematological and solid organ malignancies, autoimmune disorders, immunodeficiency states and granulomatous disorders, like tuberculosis and sarcoidosis, were recognized as predisposing conditions. The incidence rates of PML changed dramatically following the acquired immune deficiency syndrome (AIDS) pandemic in 1981 [Holman et al. 1991]. AIDS was the underlying predisposing disorder for 87% of PML cases in the USA in 1993[Selik et al. 1997], highlighting the unparalleled link between infection with human immunodeficiency virus type 1 (HIV-1) and PML. Since the introduction of highly active antiretroviral therapies, the incidence of AIDS-related PML appears to have declined [Sacktor, 2002].

The contribution of therapeutic agents to the pathogenesis of PML was not widely appreciated until the report of three natalizumab-associated cases in 2005 when PML was described in two patients with multiple sclerosis (MS) and one patient with Crohn’s disease [Kleinschmidt-Demasters and Tyler, 2005; Langer-Gould et al. 2005; Van Assche et al. 2005]. PML had not been previously reported with either MS or inflammatory bowel disorders despite the frequency with which immunosuppressive agents had been used in their treatment. Similarly, the observation of confirmed PML in three patients and suspected PML in a fourth patient who had been treated with efalizumab for psoriasis was similarly astonishing as PML had not been previously reported in psoriasis either [Korman et al. 2009; Sobell and Weinberg, 2009]. Recognition of a link between these agents which shared mechanisms of action and PML was obvious and has provided new insights into the pathogenesis of PML. A large number of other immunomodulatory drugs have been considered to increase the risk for development of PML. US Food and Drug Administration (FDA) ‘black box’ warnings for PML are found not only for natalizumab and efalizumab (since removed from the market), but also rituximab, mycophenolate mofetil, and brentuximab vedotin. However, there are substantial differences between these agents which can be classed into separate categories with respect to risk for PML. Differences include frequency with which PML is observed, the nature of the underlying disorders being treated [Garcia-Suarez et al. 2005; Carson et al. 2009] and the time from initiation of the agent to the development of PML (Table 1). Nonetheless, the risk for PML must be considered for any therapy that alters immune function.

Table 1.

Classes of agents predisposing people to progressive multifocal leukoencephalopathy (PML).

Therapeutic agent	Underlying condition previously associated with PML	Latency from initiation of therapeutic agent to the development of PML	Frequency of developing PML
Class 1	No	Long	High
Natalizumab	Multiple sclerosis and Crohn’s disease)	>8 months with peak at 24 months	Dependent on JC virus antibody status, duration of administration, and prior immunosuppressant use; range of ~1 in 1000 from 1–24 months with JC virus antibody seropositivity and no prior immunosuppressant use to >1 in 100 after 24 months of use with JC virus antibody positivity and prior immunosuppressant use
Efalizumab	Psoriasis	>3 years for efalizumab	3 of 166 receiving efalizumab for greater than 3 years
Class 2	Yes	No latency to PML development; PML develops stochastically following administration of the agent	Infrequent
Rituximab	Lymphoproliferative disorders; rheumatoid arthritis; systemic lupus erythematosus; AIDS	None	1:30,000
Mycophenolate mofetil	Solid organ transplants; systemic lupus erythematosus and other autoimmune diseases	None	?
Brentuximab vedotin	Hodgkin’s disease	None	?

AIDS, acquired immune deficiency syndrome; JC virus, John Cunningham virus.

Acquisition of JC virus

A patient must be infected with JCV to be predisposed to developing PML. Seroprevalence of JCV antibody has varied in different studies, but it has been widely accepted that seroprevalence is more than 50% in the adult population. It causes no recognized clinical symptoms at the time of initial infection, therefore, the timing and the mechanism of infection has remained elusive. Initial serological studies were based on the ability of JCV to hemagglutinate type O erythrocytes [Padgett and Walker, 1983]. Since that time, immunoassays have made serological tests more convenient [Hamilton et al. 2000] and more specific. Cross reactivity with other polyomaviruses, such as BK virus, is a potential concern, but is probably not significant [Hamilton et al. 2000]. Between the ages of 1 and 5 years, approximately10% of children demonstrate antibody to JCV and by age 10 years, it can be observed in 40–60% of the population [Taguchi et al. 1982]. Continued exposure throughout adulthood is suggested by another study that revealed seroprevalence rates of 50% in the 20–29-year-old age group, but 68% in the 68–100-year-old age group [Egli et al. 2009]. Annual rates of seroconversion of 2% have been suggested [Gorelik et al. 2010]. Studies employing an immunoassay for JCV have resulted in a broad spread of seroprevalence rates ranging from 35% [Knowles et al. 2003] to 91% [Matos et al. 2010] among adults. Seroconversion rates to JCV have exceeded 90% in some urban areas. In another study, seroprevalence rates in developed countries ranged between 50% and 60% in the adult population with MS [Bozic et al. 2011].

The mechanism of spread for the JCV remains undetermined. Respiratory or oropharyngeal transmission has been postulated based on the detection of JCV in tonsillar tissue [Monaco et al. 1998], however recent studies of these fluids in HIV-infected people and healthy controls indicate that the virus is rarely demonstrated in upper respiratory secretions and, when present, is there in very low titer [Berger et al. 2006]. Regardless of the route of infection, JCV in the urological tissues and viral shedding in urine is extraordinarily common. JCV has been demonstrated in 19% [Egli et al. 2009] to more than 60% of urine samples from immunologically normal individuals [Berger et al. 2006]. Due to its frequent occurrence in urine, perhaps it is not surprising that JCV has been detected worldwide in virtually every sample of sewage that has been examined [Bofill-Mas and Girones, 2001]. Contaminated food and water could be potential sources of infection [Bofill-Mas et al. 2001]. Importantly, the virus isolated from the urine, referred to as the archetype virus, is genetically distinct from that found in the brain tissue of people with PML. The omnipresent archetype JCV does not replicate effectively in glial tissues and must undergo genetic rearrangement to become neurotropic [Berger and Khalili, 2011].

Pathogenesis of progressive multifocal leukoencephalopathy

Despite the high prevalence rate of JCV, PML is an extremely rare disorder. This suggests that there are significant and multiple barriers to the evolution of the disorder. The following steps have been proposed in the development of PML: initial infection with JCV; establishment of a latent or persistent infection; gene rearrangement in the promoter/enhancer region of the virus to express a neurotropic form of the virus; re-expression of the virus from the sites of latency or persistence; entry of the virus into the brain; establishment of productive infection of glial tissues; failed central nervous system (CNS) immunosurveillance [Berger et al. 2009].

There remain many unanswered questions about the pathogenesis of PML. While latent JCV has been detected in tonsils, lung, spleen, bone marrow and kidney [Caldarelli-Stefano et al. 1999], controversy has surrounded its potential latency/persistence in the brain of otherwise normal individuals. B lymphocytes have been considered to play a vital role in JCV re-expression and gene rearrangement, leading to a neurotropic strain [Jensen and Major, 1999]. This genetic modification requires insertion of a 98-base-pair tandem repeat in JCV noncoding control region (NCCR). The gene rearrangement in the NCCR of the neurotropic virus allows it to bind to nuclear factor (NF)-1X binding protein, a protein that glial cells share with B cells [Houff et al. 1988; Major et al. 1990]. Either free virus or circulating JCV-infected lymphocytes cross the blood–brain barrier and infect perivascular astrocytes and oligodendrocytes. The association of PML with B-cell malignancies, such as chronic lymphocytic leukemia and Hodgkins’s disease, and disorders associated with B-cell activation, such as systemic lupus erythematosis (SLE) and AIDS, increase the conditions which place people at high risk of developing PML. Additionally, the monoclonal antibody natalizumab, which carries the highest risk among biological agents for PML development, causes hematopoietic stem cell mobilization, particularly a release of immature CD34⁺ B cells from the bone marrow [Berger and Khalili, 2011]. A summary of the proposed events leading to PML is found in Table 2.

Table 2.

Proposed steps to progressive multifocal leukoencephalopathy (PML) pathogenesis.

Initial infection with JC virus

Establishment of a latent or persistent infection

Gene rearrangement in the promoter/enhancer region of the virus to express a neurotropic form of the virus

Re-expression of the virus from the sites of latency or persistence

Entry of the virus into the brain

Establishment of productive infection of glial tissues

Failed CNS immunosurveillance.

CNS, central nervous system; JC virus, John Cunningham virus.

Class 1 therapeutic agents with risks of progressive multifocal leukoencephalopathy

Therapeutic agents that result in an unequivocal substantially increased risk of PML have been categorized as class 1. To date, two agents have been identified: natalizumab and efalizumab. These agents are associated with PML in people who have no known disorder that predisposes them to PML. Additionally, they are associated with a latency of many months to years from time of initiation to the onset of PML.

Natalizumab

Natalizumab is a recombinant humanized monoclonal immunoglobulin G4 (IgG4) antibody that binds to the α₄ subunit of the α₄β₁ subunit [Warnke et al. 2010] The α₄β₁ integrin is one of the four main integrins that are required for the firm arrest of leukocytes following their rolling and adhesion [Luster et al. 2005]. Natalizumab interferes with leukocyte α₄-mediated adhesion to its natural ligands of the extracellular matrix and endothelial lining, vascular cell adhesion molecule and fibronectin [Von Andrian and Engelhardt, 2003]. Hence, natalizumab inhibits activated lymphocytes from entering the brain. On 24 November 2004, the FDA approved natalizumab for the treatment of relapsing forms of MS. In 2005, natalizumab was withdrawn from the market after two patients with MS (treated with natalizumab and interferon β-1a) and one patient with Crohn’s disease were diagnosed with PML [Kleinschmidt-Demasters and Tyler, 2005; Langer-Gould et al. 2005; Van Assche et al. 2005]. In summer 2006, natalizumab was reintroduced as a monotherapy for relapsing-remitting MS and moderate to severely active Crohn’s disease. After its reapproval, safety and risk control trials were designed, including Tysabri Outreach: Unified Commitment to Health (TOUCH), Tysabri Global Observation Program in Safety (TYGRIS), and Crohn’s Disease – Investigating Natalizumab through Further Observational Research and Monitoring (CD INFORM). The initial estimate of PML risk was approximately 1 in 1000 people would develop PML after 17.8 months of treatment [Yousry et al. 2006]. Subsequent analyses identified three independent risks for PML development with natalizumab, namely, JCV antibody status, duration of therapy with natalizumab, and the use of immunosuppressant therapy prior to the initiation of natalizumab. The estimated risk in people with JCV antibody who are receiving natalizumab is estimated at up to 1 in 11,000, although no cases of PML have been identified in this population. The estimated incidence of PML stratified by these risk factors can be found in Table 3 [BiogenIdec, 2012].

Table 3.

Estimated incidence of progressive multifocal leukoencephalopathy (PML) stratified by risk factor (BiogenIdec 2012).

	Anti-JC virus positive
Tysabri (natalizumab) exposure	No prior immunosuppressive use	Prior immunosuppressive use
1–24 months	<1/1000	2/1000
25–48 months	4/1000	11/1000

It has been recommended that all immunomodulatory agents should be discontinued at least 3 months prior to the initiation of natalizumab [Kappos et al. 2007]. No definitive recommendations exist about the discontinuation of platform MS therapies (interferon β and glatiramir acetate) and it is quite likely that there is really no increased risk of PML with the administration of these agents. Longer washout periods may be needed for patients on immunosuppressive agents, but duration of therapy, time from last administration, and the nature of the immunosuppressant agent do not correlate with the subsequent development of PML. Rarely, PML has been reported to occur within the first 12 months of therapy, but it has not been seen prior to 8 months of therapy. The incidence rate in all natalizumab-treated patients was reported to increase after 12 months of therapy to 1.35 per 1000 and this was found to further increase to 1.76 per 1000 after 24 months. Whether longer duration of therapy would result in a further increase, a plateau or a decline in incidence remains unknown, but preliminary evidence suggests either a plateau or a slight decline. Continued surveillance of patients on natalizumab for PML is needed.

The precise explanation for an increased risk of PML in patients on natalizumab is not clear. Due to α₄β₁ antagonism, treatment with natalizumab results in impaired immune surveillance of the CNS. Reduced numbers of white blood cells including CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells, and plasma cells have been seen in the cerebrospinal fluid (CSF) of patients with MS treated with natalizumab compared with those not treated with natalizumab [Stuve et al. 2006a]. CD4⁺ cells express significantly less unbound α₄ integrin before and after natalizumab therapy compared with CD8⁺ cells, contributing to reduced migration of CD4⁺ cells across the blood–brain barrier [Stuve et al. 2006b]. There is also a reduction in the number of antigen-presenting dendritic cells and major histocompatibility complex II antigens in cerebrovascular spaces [Del Pilar Martin et al. 2008]. Natalizumab prevention of the entry of JCV cytotoxic lymphocytes into the CNS and reduction of antigen-presenting dendritic cells may predispose people to PML. Another potential contributor to the development of PML is the release of CD19⁺ CD10⁺ pre-B cells that occurs after natalizumab administration [Krumbholz et al. 2008]. These cells can be latently infected with JCV. During maturation of these B cells, transcriptional factors are activated that increase transactivation of JCV [Lindberg et al. 2008]. Rearrangement of the virus transcriptional control region leads to the evolution of neurotropic JCV. Therefore, natalizumab administration causes multiple effects on the immune system that predispose people to the development of PML.

Efalizumab

Efalizumab is a monoclonal antibody that has been approved for treatment of moderate to severe plaque psoriasis [Gordon et al. 2003; Leonardi, 2004]. It is an anti-CD 11a IgG1 antibody that binds to the α chain of CD11a and induces conformational changes in lymphocyte function-associated antigen 1 (LFA-1). LFA is the site that binds to intercellular adhesion molecule [Lub et al. 1995]. Hence, efalizumab prevents binding of T cells to endothelial molecules and blocks their passage from circulation to sites of inflammation [Vugmeyster et al. 2004]. It also inhibits activation of naïve T cells in lymph nodes and reactivation of memory T cells in response to antigen [Lebwohl et al. 2003]. Efalizumab has been shown to result in sustained improvement in psoriasis during 36 months of continuous therapy [Leonardi et al. 2008]. In animal models, blockade of CD11a on T cells caused sustained hyporesponsiveness to viral and other pathogens. This decreased responsiveness is fully reversible following efalizumab washout [Guttman-Yassky et al. 2008]. Efalizumab also reduces cutaneous dendritic cells [Lowes et al. 2005].

More than 6000 patients had been treated with efalizumab before its removal from the European and US markets in 2009. Of these, only 166 patients had received more than 3 years of treatment. There appears to be no increased risk of infection in 2335 patients receiving 12 weeks of therapy, 1115 receiving 24 weeks of therapy and 170 receiving 108 weeks of continuous therapy [Langley et al. 2005]. In 2008, three fatal cases of PML treated with efalizumab were reported. A detailed immunological study of one of these cases showed naïve T cells in peripheral blood and lack of clonal expansion in the CSF and periphery [Schwab et al. 2012]. Removal of efalizumab by plasmapheresis (PE) resulted in multiple immune changes. Polyclonal expansion of the T cells in CSF was noted with the appearance of CD8⁺ T cells 3 weeks after PE and CD4⁺ T cells 5 weeks after PE [Schwab et al. 2012]. It has been postulated that restoration of LFA function results in interaction between T-cell LFA-1 and antigen-presenting cells, restoration of memory CD8⁺ T-cell activation, and accumulation of CD8⁺ T cells in CSF [Schwab et al. 2012].

Efalizumab and natalizumab impair CNS T-cell migration and the subsequent immune responses, but some differences have been proposed in the mechanisms by which natalizumab and efalizumab lead to PML. Natalizumab causes release of progenitor bone marrow cells into the circulation, an effect not seen by efalizumab. Efalizumab predominantly impairs intrathecal antigen-presenting cells and antigen-mediated reactivation of CD4⁺ and CD8⁺ T cells, while natalizumab decreases CSF cells counts by 70% by predominantly affecting migration of T cells across the blood–brain barrier [Schwab et al. 2012].

Class 2 therapeutic agents with risks of progressive multifocal leukoencephalopathy

Therapeutic agents in class 2 are those that appear to increase the risk of PML but at significantly lower levels than class 1 agents. Most of the patients developing PML with this class of agents have either underlying conditions recognized to predispose them to PML or have been treated with other medications recognized as increasing the risk. Generally, the development of PML following their institution is a stochastic event without the need for a period of latency. The therapeutic agents in this category also carry FDA ‘black box’ warnings for PML.

Mycophenolate mofetil

Mycophenolate mofetil is a selective, noncompetitive and reversible inhibitor of inosine-5′-monophosphate (IMP) dehydrogenase. It inhibits the proliferation of T and B lymphocytes by depleting guanosine and dexoyguanosine nucleotides via inhibition of IMP conversion into guanosine monophosphate [Ransom, 1995]. Mycophenolate mofetil is approved for allograft rejection after renal, cardiac and liver transplant. It is also used for treatment of different autoimmune disorders, including SLE, myasthenia gravis and autoimmune glomerular disorders [Villarroel et al. 2009].

Cases of PML have been observed in patients taking mycophenolate mofetil but the exact contribution of this agent to the pathogenesis of PML remains unclear. All cases had underlying diseases or were on other immunosuppressive agents that predisposed them to PML. In a retrospective cohort study of nearly 33,000 renal transplant recipients in the USA, the incidence of PML among patients on mycophenolate mofetil was 14.4 cases per 100,000 versus 0 in those not receiving mycophenolate mofetil, but this did not meet statistical significance [Neff et al. 2008].

The proposed mechanism by which mycophenolate mofetil causes PML may have parallels to natalizumab. By depleting T cells, it impairs immune surveillance for JCV. Additionally, it depletes B cells and as immature B cells are re-expressed, there is a possibility of an upregulation of JCV replication and mutation to the neurotropic strain [Berger, 2010].

Rituximab

Rituximab is a chimeric monoclonal antibody directed against CD20. It depletes B and pre-B cells that express CD20. Like mycophenolate mofetil, rituximab is used for different autoimmune and hematological disorders, including autoimmune pancytopenia, SLE, rheumatoid arthritis (RA) and MS [Hauser et al. 2008; Carson et al. 2009; Gurcan et al. 2009]. It is difficult to determine the exact risk of PML with rituximab as patients identified with rituximab-associated PML had underlying immune abnormalities either related to their primary disorder or secondary to other immunosuppressants [Molloy and Calabrese, 2008]. From 1997 to 2008, 52 patients with lymphoproliferative disorders, 2 with SLE, 1 with RA, 1 each with autoimmune pancytopenia and autoimmune thrombocytopenia were identified with rituximab-related PML [Carson et al. 2009]. Data from Genentech indicate that approximately 2 million doses have been given to 1 million patients, among whom 157 cases of PML have been observed (137 with lymphoproliferative disorders, 6 with RA, 8 with SLE and 6 with AIDS). There have been no cases reported with its use in MS or other neurological disorders [Genentech, 2012].

In contrast to natalizumab and efalizumab, PML can occur any time after the administration of rituximab, but averages 5.5 months after the last rituximab dose [Carson et al. 2009]. The rate of B-cell reconstitution following rituximab administration varies from 6 months to 24 months depending on other antilymphocyte treatments, especially stem cell transplantation [Leandro et al. 2006]. Therefore, immature B cells predominate and this may contribute to transactivation and expression of a neurotropic strain of JCV. Additionally, rituximab also reduces CD3⁺ T cells in CSF [Cross et al. 2006]. One explanation for the short interval from drug administration to PML could be that affected individuals are on the verge of PML due to their underlying disorder. Administration of rituximab tips them over due to failed CNS immunosurveillance caused by the therapeutic agent [Berger and Khalili, 2012]. Similar to natalizumab and efalizumab, rituximab may affect antigen-presenting cells and T-cell regulation, though the exact effect of rituximab on T-cell function is not clear.

The incidence of PML with rituximab is considerably lower than that with natalizumab. However, the case fatality rate with rituximab-associated PML is 90% and 100% among those diagnosed with PML within 3 months of their last dose of rituximab [Carson et al. 2009]. This high rate is undoubtedly explained by the nature of their underlying disorder with a possible contribution by the irreversible nature of the immune abnormalities induced by rituximab.

Brentuximab vedotin

Brentuximab vedotin is a recently approved chimeric anti-CD30 monoclonal antibody conjugated with the antimitotic agent monomethyl auristatin E. It is approved for treatment of refractory Hodgkin’s lymphoma and relapsed or refractory systemic anaplastic large cell lymphoma. Both of these underlying lymphoproliferative disorders increase the risk of PML and it is difficult to know the degree to which the agent increased the risk of PML in the rare cases observed [Wagner-Johnston et al. 2012]. Brentuximab vedotin may decrease CNS immunosurveillance by impairing T-cell function. Alternatively, induction of the NFκβ pathway by the agent may increase JCV transcription [Wollebo et al. 2011]. However, whether brentuximab vedotin truly increases the risk of PML remains an open question.

Class 3 therapeutic agents with risks of progressive multifocal leukoencephalopathy

Therapeutic agents in this category have rarely been reported to be associated with PML and the risk of developing the disorder with their use remains uncertain.

Alemtuzumab

Alemtuzumab is a humanized monoclonal antibody against CD52, a glycoprotein that is present on most of the peripheral mononuclear cells except plasma cells. It is an approved therapy for hematological malignancies, post-transplant rejection therapy, and has been studied for MS [Fontoura, 2010]. The use of alemtuzumab causes a profound reduction of B and T lymphocytes following administration. B cells recover within 3 months, but T lymphocytes remain depleted for more than 5 years [Klotz et al. 2012]. At least three cases of PML have been described with alemtuzumab, two in patients with CLL and one in a lung transplant recipient [Martin et al. 2006; Waggoner et al. 2009]. The patient who received a lung transplant was treated for acute rejection with steroids, antithymocyte globulin, and alemtuzumab [Waggoner et al. 2009]. As with rituximab, PML has been reported with alemtuzumab in the background of either malignancy or severe immunosuppression due to coadministration of other immunosuppressants. Despite the marked lymphopenia that accompanies alemutuzumab administration, PML and opportunistic infections in general appear to be very rare and no cases of PML have been reported to date in MS cohorts who have been treated with this monoclonal antibody.

Infliximab and other tumor necrosis factor α inhibitors

Tumor necrosis factor α (TNFα) inhibitors have been widely used and the paucity of cases of PML reported with them suggests that any risk, if real, must be very small. A query of the FDA’s Adverse Event Reporting System (AERS) database for biological agents commonly employed in the treatment of psoriasis conducted in 2009 [Kothary et al. 2011] reported PML associated with adalimumab (1), etanercept (3), and infliximab (3), although these cases were not necessarily confirmed. However, in every instance the use of these agents was for conditions other than psoriasis and was confounded by the use of other immunosuppressive agents. Infliximab is a chimeric monoclonal antibody against soluble, membrane-bound TNF [Maini and Feldmann, 2002]. It is approved for the treatment of RA, psoriasis, psoriatic arthritis, Crohn’s disease, ulcerative colitis, and ankylosing spondylitis. One PML case has been reported with infliximab in a patient with RA [Kumar et al. 2010]. PML developed after 3 years of intake of infliximab, methotrexate and steroids. Infliximab suppresses TNFα, which decreases interleukin-6 and interleukin-1, thus reducing inflammation. A decrease in cell recruitment occurs through reduced expression of CD3⁺, CD68⁺, vascular cell adhesion molecule 1, intercellular adhesion molecule, and E-selectin. Dendritic cell-mediated T-cell activation is also reduced [Gottlieb et al. 2005]. This significant immune modulation, particularly T-cell activation, might be implicated in the pathogenesis of PML.

Fludarabine

Fludarabine is a purine nucleoside that has been used as a chemotherapeutic agent for the treatment of hematological malignancies. It is cytotoxic against resting and dividing cells [Keating et al. 1989]. In resting cells, it inhibits DNA synthesis; in dividing cells, it prevents the DNA repair process and induces apoptosis. Fludarabine is used for the treatment of chronic lymphocytic leukemia, which is the most common hematological malignancy associated with PML. Hence, association of fludarabine and PML has been questioned in the past. In most reported cases PML developed after fludarabine therapy [Gonzalez et al. 1999; Leonard et al. 2002; Saumoy et al. 2002; Lejniece et al. 2011], therefore it is believed that immunosuppression resulted in JCV activation and impaired clearance.

Fumaric acid

Fumaric acid esters (FAEs) are a group of related compounds that have been used for the treatment of psoriasis since 1959. Several lines of evidence suggested immunomodulatory effects for FAEs. FAEs have been shown to reduce the number of T cells, particularly, CD4⁺ T cells, induce apoptosis of T cells [Treumer et al. 2003], and inhibit translocation of NFκB into the nucleus. Disruption of the NFκB pathway leads to decreased expression of NFκB genes that are involved in regulating various inflammatory cytokines, chemokines, and adhesion molecules [Stoof et al. 2001]. NFκB has been speculated to inhibit B-cell antiapoptotic protein, resulting in increased apoptosis of B cells, though no definitive studies suggest a direct effect of FAEs on B cells [Moharregh-Khiabani et al. 2009]. A phase II clinical trial in patients with relapsing-remitting MS receiving dimethylfumarate (BG-12) showed a significant reduction in the number of gadolinium-enhancing lesions after 24weeks [Kappos et al. 2008]. Phase III trials of BG-12 have demonstrated safety and efficacy in relapsing-remitting MS [Moharregh-Khiabani et al. 2009].

Recently, a fumaric-acid-related case was reported in Germany in a 74-year-old man who developed pathologically proven PML after 3 years of treatment with fumaric acid for psoriasis. Whether there were other predisposing risks for PML was not precisely stated [Ermis et al. 2011]. All other cases of PML reported with fumaric acid from a German database of Fumaderm (Fumedica/Hermal) have had other risk factors for its occurrence, such as, sarcoidosis and treatment with efalizumab.

Risk mitigation strategies

Different strategies have been suggested to mitigate the risk of PML in patients treated with drugs that increase the risk. As JCV exists in at least 50% of the adult population as a latent or persistent infection, serological tests should be done to determine whether the individual has been previously exposed to the virus. Evidence that PML is the consequence of reactivation of a latent infection includes the presence of the IgG antibody to JCV [Weber et al. 2001], the rarity of PML in children [Berger et al. 1992], the demonstration of genetically identical JCV isolates from the blood and from other tissues months to years before the onset of PML and their isolation from the brain [Fedele et al. 2003], and demonstration of the JCV antibody in all patients to date who have developed natalizumab-associated PML and had stored blood samples available for testing. A commercially available test for JCV antibody is available and should be performed in patients who are being started on natalizumab and other therapies carrying a substantial risk of PML. If a patient is seronegative, repeat determinations at regular intervals (perhaps 6 months) are warranted.

Experience with natalizumab-associated PML has demonstrated that early detection of the disease is an important determinant of survival and outcome. This probably reflects rapid removal of the agent with plasmapheresis and restoration of normal immune function. In radiographically isolated PML, that is, detected by magnetic resonance imaging (MRI) before the recognition of clinical symptoms [Langer-Gould et al. 2005; Linda et al. 2009], periodic cranial MRIs are warranted. The use of fluid-attenuated inversion recovery images may prove sufficient in detecting suspicious lesions without resorting to a standard set of sequences. The optimal interval for repeat MRI has not been determined, but 6-month intervals once the patient has exceeded a time period of 12 months on natalizumab would appear to be sufficiently prudent.

Other strategies to reduce the risk of PML in patients on these therapeutic agents, in particular natalizumab, such as reducing the dose or frequency of administration, offering drug holidays, or the coadministration of serotonin receptor blockers, remain unsupported by any scientific studies to date. A high index of suspicion for PML needs to be maintained for individuals at risk to identify the disease as early as possible.

Conclusion

PML is a rare demyelinating disorder that is seen in the setting of a disturbed immune system. AIDS, hematological malignancies and many autoimmune conditions predispose people to the development of PML. PML has been observed with the use of new biological agents in the treatment of these disorders. Efalizumab and natalizumab are unique monoclonal antibodies that have caused PML in patients with psoriasis (efalizumab) and MS (natalizumab), conditions that do not increase the risk of PML development. The occurrence of PML with these and other therapeutic agents has provided insight into and spurred the investigation of the pathogenesis of PML. In time, a better understanding of the specific risks associated with these newer therapies and refined strategies for PML risk reduction will undoubtedly be attainable.