TABLE 3 Summary of agents predisposing to PMLAgents Mechanisms of action Possible explanation forincreased risk of PMLNatalizumabα4β1 and α4β7 integrinantibodies↓ JCV-specific CTLs traffickinginto CNS; ↓ CNSperivascular dendritic cellsfor antigen processing; ?↑neurotrophic JCV expressionEstimatedrisk of PML1.51:1,000patientstreated*Unique predispositionfor PML aYesEfalizumab Anti CD11a antibody Blockade of co-stimulatorymolecules on T-cells; ↓ JCVspecificCTLs trafficking intoCNSRituximab Anti CD20 antibody ↑ JCV expression with recoveryof B-cell population;↓ B-cell antigen presentation1:500 6 Yes1:25,000** NoPML = Progressive multifocal leukoencephalopathy; CTLs = cytotoxic T-lymphocytes; ↓ = decreasing; ↑ = increasing; ?↑=possiblyincreasing; a = risk of development of PML with the drug in the absence of underlying disorders that increase its risk. (Adaptedfrom Berger JR. Progressive multifocal leukoencephalopathy and newer biological agents. Drug Saf. 2010;33:969-83.)* Data on file. Tysabri safety update. Weston, MA: Biogen Idec; June 2011. ** Data extracted from Clifford DB, Ances B, Costello C,et al. Rituximab-associated progressive multifocal leukoencephalopathy in rheumatoid arthritis. Arch Neurol. 2022 (in press).blood circulation, 14,17 it is plausible to thinkthat immature JCV-infected leukocytes maybe released from the bone marrow after thesemonoclonal antibody treatments, inducingviremia. 18 Indeed, a small (n = 19) study demonstratedthe increased prevalence of JCV inurine, plasma, and peripheral-blood mononuclearcells (PBMCs) after the initiation of natalizumabtreatment among MS patients. 19 However,a larger (n = 1397) study by Rudick andcolleagues 20 failed to confirm these findings,and another study did not detect JCV DNA inthe PBMCs and CD34+ hematopoietic precursorcells in natalizumab-treated patients, arguingagainst the JCV infection of CD34+ cells andits mobilization as possible explanation of PMLdevelopment after natalizumab treatment. 17Natalizumab binds with adhesion moleculesα4β1 and α4β7 integrins, 21 whereas efalizumabbinds to the alpha chain (CD11a) of the leukocytefunction–associated antigen (LFA-1), 22 inhibitsthe migration of lymphocytes across theblood-brain barrier (BBB), and may possiblycreate compartmentalized cell-mediated immunedeficiency in the CNS, facilitating localJCV reactivation and PML development.Rituximab, which was first approved bythe US Food and Drug Administration (FDA)in 1997, was found to be associated with PML.As of 2009, 57 HIV-negative patients who weretreated with rituximab and other agents developedPML. 23 Fifty-five of these PML patientshad lymphoid malignancies, while only twohad lupus and one each had rheumatoid arthritis,idiopathic autoimmune pancytopenia,or immune thrombocytopenia purpura. However,most of these patients were treated withseveral other classes of immunosuppressiveagents, and the exact role of rituximab in PMLdevelopment is unclear. 8 Rituximab binds withCD 20 expressed on B-lymphocytes and depletestheir population in the blood 23 and alsopossibly releases the JCV-infected immatureB-lymphocytes.DiagnosisSince the clinical features of PML vary accordingto the localization of the demyelinating lesionsand are nonspecific, diagnosis of PMLrequires strong clinical vigor supported bylaboratory tests. Methods for diagnosing PMLinclude polymerase chain reaction (PCR) detectionof JCV DNA in the cerebrospinal fluid(CSF) and detection of viral DNA or proteinsby in situ hybridization or immunohistochemistryon a brain biopsy sample. 8 In addition,S20 July 2011 • Clinical Reviews of JCV and PML
characteristic brain changes on magnetic resonanceimaging (MRI) are useful. 24 PCR detectionof JCV in CSF has been a gold standard forconfirming PML diagnosis; however, becauseof cART its sensitivity in AIDS-associated PMLmay be low, ranging from 42% to 81%. 25-27 Whilethe PCR method is also useful in diagnosingmonoclonal antibodies- and IRIS-associatedPML, PCR may fail to detect low JCV in CSF ofthese individuals because JCV replication maybe held by a competent immune system in theformer and recovering immune system in thelatter, and a brain biopsy may be necessitated. 8Since only a small fraction of monoclonalantibody–treated individuals are likely todevelop PML, which can lead to long-termneurological sequelae or death, stratifying patientsaccording to risk of PML development isof paramount importance. Indeed, a series ofrecent studies have determined three factorsthat are useful in stratifying monoclonal antibody-treatedMS patients according to theirrisk of PML development. These factors areanti-JCV antibody–positive status, prior immunosuppressantuse, and natalizumab treatmentduration.In a recent editorial, Tyler 5 raised the questionof whether risk can be reduced in patientsreceiving biological immunomodulatorytherapies. “The ability to stratify risk moreaccurately in specific subpopulations wouldobviously be of tremendous utility in guidingtherapeutic decisions,” he wrote. “Diseaseprevention and risk reduction are even morecritical because of the absence of any specificantiviral therapy of proven benefit in the treatmentof PML.”Even though no subjects in their study developedPML, Chen and colleagues 19 suggestedthat surveillance of blood or urine may be usefulin identifying patients potentially at risk ofdeveloping PML after natalizumab treatment.However, a larger prospective study by Rudickand colleagues 20 testing for JCV DNA using acommercially available quantitative PCR assayas well as a more sensitive quantitative PCRassay developed at the National Institutes ofHealth found no difference in the prevalenceof JCV DNA–positive plasma in placebo-treatedor natalizumab-treated patients, nor in theprevalence of similarly positive urine specimensat baseline and at 48 weeks. Furthermore,JCV DNA was not detected prior to diseaseonset in the blood of the five patients whodeveloped PML during the study. Rudick et al 20concluded that “measuring JCV DNA in bloodor urine with currently available methods isunlikely to be useful for predicting PML riskin natalizumab-treated MS patients,” a findingwith which Tyler concurred. 5Pointing to detection of virus-specific antibodiesas the traditional method for establishingpast exposure to viral infection, Tyler 5went on to observe that “A highly sensitive andspecific JCV antibody assay would theoreticallyenable the population to be divided intothose at risk for developing PML (antibodypositive), and those not at risk (antibody negative).”While early studies of JCV serostatus indicatedthat the primary infection happenedin infancy and childhood and that, by adulthood,sero prevalence rates as high as 80%were reported, these findings may have beenconfounded by use of a nonspecific hemagglutination-inhibitionassay, a less accurate butwidely available assay method of that time, aswell as by potential antibody cross-reactivitybetween antibodies against JCV and other humanpolyomaviruses.Two recent studies of note have useddifferent forms of enzyme-linked immunosorbentassay (ELISA) techniques. 7,28 In anevaluation of MS patients treated with natalizumabusing a two-step ELISA to detect antibodiesto the JCV-like particles, Gorelik andcolleagues 28 demonstrated that 100% (17 outof 17) of sera from natalizumab-associatedPML patients obtained 16 to 180 months priorto PML development were JCV-seropositive,whereas only 53.6% of non-PML MS patientstreated with natalizumab were anti-JCV–seropositive.The investigators interpreted theirresults as warranting further research on theclinical utility of the assay as a potential PMLrisk stratification tool for MS patients. Similarly,another ELISA study using baculovirus expressedJCV VP1 showed that 65% of sera from214 MS patients treated with natalizumab andClinical Reviews of JCV and PML • July 2011 S21