Cellular and Humoral Immune Response inProgressive Multifocal Leukoencephalopathy

Frank Weber, MD,1 Claudia Goldmann,2 Marcus Kramer,2 Franz Josef Kaup, MD,3 Marcus Pickhardt, PhD,4

Peter Young, MD,5 Harald Petry, PhD,2 Thomas Weber, MD,4 and Wolfgang Luke, PhD2

Progressive multifocal leukoencephalopathy (PML) is a fatal, demyelinating disease caused by JC virus (JCV) in patientswith severe immunosuppression. We studied the JCV-specific cellular and humoral immune response in 7 healthy donors(HD), 6 human immunodeficiency virus-1 (HIV-1)-infected patients without PML (HIV), 4 HIV-1-negative patientswith PML (PML), and 8 HIV-1-positive patients with PML (HIV/PML). As antigens, recombinant virus-like particles ofthe major structural protein VP1 (VP1-VLP) of JCV, tetanus toxoid (TT), or the mitogen phytohemagglutinin (PHA)were used. Proliferation of peripheral blood mononuclear cells (PBMC) after stimulation with the VP1-VLP was signif-icantly suppressed in PML and HIV/PML patients compared to HD. After antigen stimulation the production ofinterferon-g (IFN-g) was reduced in PML, in HIV/PML, and in HIV patients. The production of interleukin-10 (IL-10),however, was elevated in HIV/PML patients. Neither proliferation nor cytokine production correlated with the presenceof JCV DNA in PBMC. The immunoglobulin G serum antibody titer to the VP1-VLP was slightly elevated in HIV,elevated in PML, and highly elevated in HIV/PML patients compared to HD. The development of PML appears tocoincide with a general impairment of the Th1-type T-helper cell function of cell-mediated immunity.

Ann Neurol 2001;49:636–642

Progressive multifocal leukoencephalopathy (PML) wasfirst described in 1958.1 PML is a demyelinating dis-ease of immunocompromised patients caused by thehuman polyomavirus JC virus (JCV). Once PML was arare disorder, but the incidence has risen dramaticallyas a result of the worldwide AIDS pandemic. The ac-quired immune deficiency syndrome (AIDS) is the un-derlying immunosuppression in about 85% of PMLpatients and about 4–6% of all patients with AIDSmay develop PML.2,3 The disease is also observed inpatients with neoplastic diseases and in immunocom-promised patients after transplantation or treatmentwith immunosuppressants.3,4 Very rarely, PML is seenin patients without any signs of immunodeficiency.5

PML results from a lytic infection of oligodendrocytesby JCV, which causes death of oligodendrocytes anddemyelination. Clinical symptoms of PML are nonspe-cific and include motor weakness, cognitive changes,visual deficits, and cranial nerve palsies.6 By magneticresonance imaging (MRI), multifocal demyelinating le-sions are most often found at the gray matter–whitematter junction. Detection of JCV DNA in the cerebro-spinal fluid (CSF)7 or of an intrathecal JCV-specific an-tibody synthesis8 is used to confirm diagnosis of PML.

Whereas many studies address the molecular biologyof JCV, studies on the cellular and humoral immuneresponse are rare.9,10 This is mainly due to the limitedavailability of viral antigen based on the difficulties ofpropagating JCV in cell culture. We describe a system-atic analysis of the humoral and cellular immune re-sponse to JCV. To obtain sufficient amounts of JCV-related antigen, we expressed the major structuralprotein VP1 of JCV as virus-like particles (VP1-VLP).11 These VP1-VLP were used to investigate theproliferation of peripheral blood mononuclear cells(PBMC) as well as the cytokine production in sevenhealthy donors (HD), six HIV-1-infected patientswithout PML (HIV), eight HIV-1-infected patientswith PML (HIV/PML), and four HIV-1-negative pa-tients with PML (PML). In addition, JCV-specific se-rum immunoglobulin G antibodies were assessed usingVP1-VLP as antigen.8,12

Patients and MethodsPatients and ControlsUp to 40 ml EDTA blood was obtained from 7 HD, 6 HIV,4 PML, and 8 HIV/PML patients after informed consentwas obtained. There was no statistically significant difference

From the 1Department of Neurology, University of Gottingen; De-partments of 2Virology and Immunology and 3Experimental Pathol-ogy, German Primate Center, Gottingen; 4Department of Neurology,Marienkrankenhaus, Hamburg; and 5Department of Neurology, Uni-versity of Munster, Germany.

Received Dec 15, 1999, and in revised form Aug 11. Accepted forpublication Nov 27, 2000.

Published online 2 April 2001.

Address correspondence to Dr Weber, Section of Neurology, MPIof Psychiatry, Kraepelinstrasse 2–10, 80804 Munich, Germany. E-mail: fweber@mpipsykl.mpg.de

636 © 2001 Wiley-Liss, Inc.

with respect to CD41/CD81 ratio and HIV-1 virus loadbetween HIV and HIV/PML patients (Table 1). Among the4 PML patients, 2 suffered from chronic lymphatic leukemiaand 1 from immunosuppression after organ transplantation.In the fourth patient, no underlying immunosuppressive ill-ness could be detected. PML was clinically suspected basedon the occurrence of neurological deficits and typical lesionson magnetic resonance imaging (MRI).3,13 Diagnosis wasproved by the detection of JCV DNA in cerebrospinal fluid7

or by the demonstration of a JCV-specific intrathecal anti-body response.8,12 In 2 HIV-1-negative PML patients, thediagnosis was proved by brain biopsy. PBMC of HD wereincluded in the study, to investigate the proliferative capacityand the cytokine production in healthy volunteers. Thisstudy was approved by the ethics committee of the Georg-August-University Gattingen and the institutional reviewboard of the Arztekammer, Hamburg, Germany.

Template PreparationFrom patients who donated 40 ml EDTA blood or more,sufficient amounts of PMBC could be isolated to analyze thepresence of JCV DNA in PBMC by polymerase chain reac-tion (PCR). These were 4 HIV, 3 PML, and 5 HIV/PMLpatients. In addition, 3 HD with and 3 HD without JCV-specific antibodies were investigated. DNA was isolated from106 PBMC. Cells were lysed by 1% sodium dodecyl sulfate(SDS) and incubated in the presence of proteinase K (300mg/ml) overnight at 56°C. Thereafter, an equal volume ofbuffer-saturated phenol was added and mixed, and DNA wasextracted by phenol-chloroform extraction. Finally, the DNAwas precipitated by the addition of 0.1 volumes of a satu-rated NaCl solution followed by an equal volume of isopro-

panol for 1 hr at –20°C. DNA was pelleted at 20,000g for15 minutes; DNA pellets were washed repeatedly with 70%ethanol, dried, and finally suspended in 100 ml H2O.

PCR AmplificationNested PCR was carried out using 1 mg of purified DNA aspreviously described,7 with some modifications. The firstround of PCR consisted of 10 cycles with the primer pairJC4284/JC4961 flanking a 656-base-pair fragment of thelarge T gene. The sequence of primer JC4284 in 59 to 39orientation was from position 4,284 to position 4,527 TTTTTA GGT GGG GTA GAG TGT TG and primer JC4961from position 4,982 to 4,961 TGG AGC TTA TGG ATTTAT TAG G. The second round of PCR was carried outwith 4 ml of the first round for 30 cycles using the primerpair JC36/JC39 spanning a 389-base-pair fragment locatedin the large T gene region.6 In 59 to 39 orientation theprimer JC36 has the sequence AGA TCC CTG TAG GGGGGT GTC TCC from position 4,505 to position 4,527 andprimer JC 39 from position 4,894 to position 4,874 the se-quence AAT GCA AAG AAC TCC ACC CTG ATA AAGGTG G. Each amplification round was initiated at 95°C for2 minutes 30 seconds. Cycles of first-round PCR consistedof 95°C for 30 seconds, 55°C for 1 minute, and 72°C for 1minute. Second-round PCR consisted of 95°C for 30 sec-onds, 60°C for 1 minute, and 72°C for 1 minute. Bothrounds were concluded with an extension cycle of 72°C for 5minutes. By this PCR, at least 10 copies of target JCV DNAmolecules were detected in a background of a JCV-negativecell extract from 106 cells of a permanent human T cell line,which was spiked with 106 to 0 JCV DNA molecules.

Table 1. Characteristics of HIV-1-Infected Patients

PatientsCD4(Cells/ml)

CD4/CD8(Ratio)

HIV-1 Virus Load(RNA Copies/ml) Antiretroviral Therapy

HIVRO 660 0.50 5,500 NoneHW 220 0.60 1,600 Azidothymidin, lamivudinLK 160 0.32 0 Azidothymidin, lamivudin,

saquinavirKR 300 nd 0 Azidothymidin, lamivudin,

saquinavirMM 22 0.02 560,000 NoneGE 204 0.38 24,000 None

HIV/PMLMF 443 0.50 158,000 NoneVP 373 0.40 nd IL-2, lamivudin, stavudinWM 346 0.14 14,500 Dideoxyinosin, saquinavir,

stavudinBT 60 0.19 600 Azidothymidin, lamivudin,

saquinavirHA 250 0.24 nd Azidothymidin, indinavir,

lamivudinUM nd nd nd Fivefold combinationBK nd nd nd Azidothymidin, indinavir,

lamivudinOR 77 0.07 3,700 Fivefold combination

HIV 5 HIV-1-infected patients without PML; HIV/PML 5 HIV-1-infected PML patients. There was no difference between HIV-1-infectedpatients with or without PML with respect to CD4/CD8 ratio and HIV-1 virus load (p 5 0.428 and p 5 0.699, respectively).

Weber et al: Immune Response in PML 637

Analysis of PCR ProductsSecond-round PCR products were purified by a PCR puri-fication kit according to the manufacturers procedure (Qia-gen, Hilden, Germany). For proving JCV specificity, the pu-rified DNA was digested with the restriction endonucleaseRsa I according to the instructions of the manufacturer (Am-ersham Pharmacia Biotech, Freiburg, Germany) and sepa-rated on a 2% agarose gel. In case of JCV specificity, twofragments of about 100 and 300 base pairs will arise afterrestriction enzyme digestion.

AntigensThe major structural protein VP1 of JCV was cloned andexpressed in insect cells as previously described.11 Recombi-nant VP1 spontaneously assembled to virus-like particles re-sembling empty JCV capsids.11 Tetanus toxoid (TT) was akind gift of Dr Blackkolb (Behringwerke, Marburg, Germa-ny), and phytohemagglutinin (PHA) was purchased fromSigma (Deisenhofen, Germany).

Preparation of PBMC and Proliferation AssayPeripheral venous blood was separated within 24 hr byFicoll-Hypaque gradient centrifugation. PBMC were washedtwice with phosphate-buffered saline (PBS) and seeded intoround-bottomed 96 well culture plates (Nunc, Roskilde,Denmark) at 2 3 105 cells per well in a final volume of 100ml CG medium (Vitromax, Selters, Germany) supplementedwith 2 mM L-glutamine, 103 U/ml penicillin and 103 U/mlstreptomycin (Gibco, Eggenstein, Germany). VP1-VLP wereadded at a final concentration of 10 mg/ml. Parallel sets ofwells containing TT (10 mg/ml), PHA (5 mg/ml), or me-dium alone were used as controls. Cells were maintained at37°C in a humidified atmosphere (5% CO2).

Four days after antigen stimulation, PBMC were labeledwith 0.1 mCi [3H]thymidine (Amersham, Braunschweig,Germany) and harvested 18 hr later. [3H]thymidine incor-poration was measured with a scintillation counter (Tri-Carb4000 Series; Packard, Frankfurt, Germany). All determina-tions were set up at least in triplicate. The stimulation index(SI) was calculated as follows: counts per minute (cpm) ofantigen-stimulated culture/cpm of culture with medium.

Determination of CytokinesPBMC cultures were set up as described above for the pro-liferation assay. After 2, 3, and 4 days the content of 3 wellswas pooled and pelleted (2,200g, 5minutes). The supernatantwas stored immediately at –80°C. The Th1-associated cyto-kine interferon-g (IFN-g) and the Th2-associated cytokineinterleukin-10 (IL-10) were determined by a sandwichenzyme-linked immunosorbent assay (ELISA). Antibodiesand standard were purchased from Biosource (Ratingen,Germany). The ELISA was performed according to the in-structions of the manufacturer, with minor modifications.All samples were determined in duplicate. The detection lim-its of IFN-g and IL-10 were determined at 3.4 pg/ml and2.4 pg/ml, respectively.

VP1-Specific Antibody ELISAThe titer of VP1-specific immunoglobulin G antibodies insera of patients and HD was determined as described previ-

ously.8 In brief, ELISA plates (Nunc) were coated with 100ml VP1-VLP (1 mg/ml) and incubated with serial dilutionsof sera. Human IgG was detected by a POD-conjugated goatantihuman IgG (Dianova, Hamburg, Germany). Resultswere expressed as arbitrary units, which were standardizedusing a human serum pool.

Statistical AnalysisThe Mann-Whitney test was used to calculate significance ofresults. A p value , 0.05 was regarded as significant.

ResultsOptimization of PBMC ProliferationIn pilot experiments, the dose response of PBMC toincreasing amounts of VP1-VLP was investigated in 4members of the HD group. In the presence of 0.1–10mg/ml VP1-VLP, a dose-dependent proliferation wasdemonstrated in all HD. A representative result isshown in Figure 1. A similar dose response was ob-tained with TT as the stimulating antigen (data notshown). In the following experiments, VP1-VLP andTT were used at a concentration of 10 mg/ml. Furtheranalysis showed that not only CD41 but also CD81

cells were able to proliferate in response to the VP1-VLP (data not shown).

Proliferation of PBMC Obtained from HD andHIV, PML, or HIV/PML PatientsFigure 2A shows that proliferation after stimulationwith the VP1-VLP was highest in the HD group.Compared to the case in the HD group, the prolifer-ation was significantly suppressed in both the PML(p 5 0.004; Fig 2A) and the HIV/PML groups (p 50.0002; Fig 2A). In the HIV/PML group, the SI valueswere lower than in the HIV group. However, this dif-ference was not statistically significant (p 5 0.21; Fig

Fig 1. The proliferation of peripheral blood mononuclear cellsagainst VP1-VLP is dose-dependent. A representative result ofa healthy donor (HD) is shown. Values represent mean 6SD. All assays were performed in quadruplicate.

638 Annals of Neurology Vol 49 No 5 May 2001

2A). Stimulation of PBMC with TT yielded similarresults. The HD group showed the highest prolifera-tion (Fig 2B). Whereas the response to TT in the PMLgroup was significantly lower than in the HD group(p 5 0.012), the difference between the HIV groupand the HIV/PML group was small and not statisti-cally significant (p 5 0.19). After stimulation with

PHA, the PBMC proliferation in the HD group washigher than in all other groups (Fig 2C), but differ-ences between the groups were not statistically signifi-cant (PML and HD: p 5 0.57; HIV and HIV/PML:p 5 0.6).

Cytokine Production of PBMC from HD and HIV,PML, or HIV/PML PatientsIFN-g secretion after stimulation with VP1-VLP, TT,and PHA was highest in the HD group (Fig 3A–C). Amarginal IFN-g production was found in all othergroups after stimulation with VP1-VLP or TT (Fig3A,B). After stimulation with PHA, the IFN-g produc-tion was much lower in the HIV, PML, and HIV/PML groups compared to the HD group (Fig 3C). Se-cretion of IL-10 by PBMC obtained from HIV/PMLpatients in response to VP1-VLP or TT was clearly el-evated compared to all other groups (Fig 3D,E). Afterstimulation with PHA, however, there was a slight dif-ference in favor of the HD and PML groups in com-parison to both HIV-1-infected groups (Fig 3F).

Fig 3. Peripheral blood mononuclear cells (PBMC) of healthydonors (HD), progressive multifocal leukoencephalopathy(PML) patients without HIV-1 infection (PML), HIV-1-infected patients without clinical signs of PML (HIV), andHIV-1-infected PML patients (HIV/PML) were stimulatedwith VP1-VLP (10 mg/ml; A,D), TT (10 mg/ml; B,E), andPHA (5 mg/ml; C,F). Production of IFN-g and IL-10 wasdetermined by ELISA on days 2, 3, and 4 after stimulation.Values are given as mean over 3 days for each patient. Inaddition, the median of each group is indicated.

Fig 2. Proliferation of peripheral blood mononuclear cells(PBMC) of healthy donors (HD), progressive multifocal leu-koencephalopathy (PML) patients without HIV-1 infection(PML), HIV-1-infected patients without clinical signs of PML(HIV), and HIV-1-infected PML patients (HIV/PML) in thepresence of VP1-VLP (10 mg/ml; A), TT (10 mg/ml; B), andPHA (5 mg/ml; C) is shown. The difference in proliferationto VP1-VLP (A) and TT (B) between HD and PML is sta-tistically significant ( p 5 0.004 and p 5 0.012, respectively).The difference in proliferation between HIV and HIV/PML isstatistically not significant ( p 5 0.21 and p 5 0.19, respec-tively). Proliferation was determined by [3H]thymidine incor-poration. Values are given as stimulation index and median.

Weber et al: Immune Response in PML 639

JCV DNA Detection in PBMC and JCV-SpecificHumoral Immune ResponsePBMC JCV DNA was detected in 50% of HD, in75% of HIV patients, in 66% of PML patients, and in60% of HIV/PML patients (Table 2). All PCR prod-ucts were demonstrated to be JCV-specific by restric-tion endonuclease digestion (data not shown). In theHD group, the presence of JCV DNA in PBMC cor-related with the occurrence of JCV-specific antibodies.No correlation, however, was found in any other group(Table 2). The highest VP1-specific antibody titerswere found for HIV/PML patients, who showed a verylow proliferation in response to VP1-VLP. No correla-tion of the SI values after PBMC stimulation withVP1-VLP, TT, or PHA and the presence of JCV DNAin the PBMC was found in any of the groups. Thehigh VP1-specific antibody titer in the HIV/PMLgroup is in agreement with the IL-10 productionfound in this group. However, there was no significantdifference between the antibody titers of the PML andthe HD groups or between the HIV and the HIV/PML groups (p 5 0.11 and p 5 0.06, respectively).

DiscussionInvestigations of the immunopathogenesis of PML arelimited to humans, because no animal model for PMLis known.14 Only a few studies have addressed the cel-lular and humoral immune response to JCV inPML.9,10,15 We analyzed the JCV-specific cellular andhumoral immune response in HIV-1-infected patientswithout or with PML and in PML patients withoutHIV-1 infection. HD were included in the study toanalyze the normal situation at baseline. The presentinvestigation became feasible with the availability ofsufficient amounts of the recombinant major structuralprotein VP1 of JCV.11 The recombinant VP1 formedVP1-VLP, which exhibits the same structural and func-tional properties as natural JCV capsids.11 In an earlierstudy, we demonstrated a highly reduced proliferative

response to purified natural JCV particles in a PMLpatient compared to HD.16 This observation is in linewith the findings reported here for VP1-VLP, whichappeared to be a very useful antigen for the assessmentof the JCV-specific cellular and humoral immune re-sponse.

The present study demonstrates that a pronouncedJCV-specific proliferative response in HD is obviouslyone of the important mechanisms to control JCV. Asignificantly lower JCV-specific proliferation of PBMCcompared to HD was found in PML and HIV/PMLpatients. Furthermore, the proliferative response ofPBMC to TT and PHA was also impaired in PMLpatients compared to HD. Our observation of a re-duced lymphocyte proliferation in response to PHAstimulation is in line with earlier reports10,16 and theprevious description of the absence of a delayed-typehypersensitivity to various antigens in PML pa-tients.10,15

In parallel with the suppressed proliferation, the pro-duction of the Th1-type cytokine IFN-g, which is as-sociated with an antiviral cellular immune response,was strongly reduced in PML, HIV/PML, and HIVpatients after stimulation with VP1-VLP, TT, andPHA. In contrast, the production of the Th2-type cy-tokine IL-10, which is associated with an antiviral hu-moral immune response, was elevated in response toVP1-VLP and TT in HIV/PML. Neither the prolifer-ative capacity nor the cytokine production was corre-lated to the JCV DNA detected in the PBMC. Thefrequency of JCV DNA was in line with several stud-ies.17 The relatively high percentage of JCV DNA de-tected in the PBMC of patients in the HIV group wassimilarly correlated with the low CD41 T cell count,as previously reported.18

The elevated titer of VP1-specific IgG antibodies inPML and HIV/PML patients compared to HD andHIV patients provides further evidence for an antiviralhumoral immune response. These findings support the

Table 2. JCV-Specific Humoral Immune Response and Detection of JCV DNA in PBMC by PCR

HD PML HIV HIV/PML

Donor Titers PCR Donor Titers PCR Donor Titers PCR Donor Titers PCR

WL 4,800 nd ED 0 1 RO 6,800 1 MF 8,400 ndMK 1,664 1 GH 60,000 2 HW 0 2 VP 316,000 ndSZ 0 2 PW 3,500 1 LK 44,000 nd WM 7,584,000 2SP 0 2 KR 43,200 1 BT 353,600 1EM 4,500 1 MM 0 nd HA 2,480 2CG 0 2 GE 0 1 UM 87,000 1KL 17,500 1 BK 150,000 nd

OR 0 1Median: 1,664 33,550 3,400 118,700

HD 5 healthy donors; PML 5 PML patients without HIV-1 infection; HIV 5 HIV-1-infected patients without PML; HIV/PML 5 HIV-1-infected PML patients; nd 5 not done; PCR 5 polymerase chain reaction; 1 5 .10 JCV molecules; 2 5 ,10 JCV molecules; PBMC 5peripheral blood mononuclear cells. Antibody titers were expressed as arbitrary units.

640 Annals of Neurology Vol 49 No 5 May 2001

hypothesis of an impaired cellular immune responseleading to the reactivation of JCV and the develop-ment of PML. Although Willoughby et al.10 describeda specific deficiency in the cellular immune response toJCV in PML patients, our more detailed study doesnot support a JCV-specific immunodeficiency. This iscorroborated by the observation of a reduced PBMCproliferation and suppressed production of IFN-g afterstimulation with TT, PHA, and VP1-VLP. Our resultsindicate a general loss of Th1-type T-helper functionsin PML patients as evidenced by the reduction of pro-liferation and IFN-g production in the presence of allthree stimulants. However, to support this observation,more Th1-type cytokines (interleukin-2, tumor-necrosis-factor-a, and lymphotoxin) should be investigated infurther studies. This hypothesis is in agreement with sev-eral reports emphasizing the importance of IFN-g in thecontrol of a number of virus infections.19 In addition,the effect of highly active antiretroviral therapy(HAART) argues against a selective JCV-specific immu-nodeficiency. It has recently been shown that HAARTsignificantly prolongs the survival of HIV-1-infectedPML patients.12,20–27 The positive effect of HAARTappears to be due to the reconstitution of the immunesystem and correlates with an increase of the number ofCD41 T cells.23–25

In contrast to the diminished proliferation ofPBMC, JCV-specific serum antibodies are elevated inmost PML patients, in particular in the HIV/PML pa-tients. Obviously, the B cell response against JCV isnot impaired in most PML patients and may be T cell-independent. Such a T cell-independent humoral im-mune response occurs after infection with the mousepolyomavirus28 and provides resistance to acute mor-tality.29 The finding of a high antibody titer in HIV/PML patients together with an elevated IL-10 produc-tion after stimulation with VP1-VLP supports afunctional Th2-type T-helper response. It may reflect ashift from a Th1-type to a Th2-type immune responsein these patients. Recent evidence also suggests an as-sociation of prolonged survival of PML patients receiv-ing HAART with the development of a VP1-specificintrathecal humoral immune response.12 The demon-stration of a JCV-neutralizing capacity of serum or in-trathecal VP1-specific antibodies would elucidate therole of the humoral immune response in the develop-ment of PML. Because the VP1-VLP carry authenticneutralizing epitopes of JCV,11 they represent an idealantigen to perform such a detailed study for JCV-specific neutralizing antibodies.

Furthermore, recent findings of a comparable inci-dence of PML in patients receiving HAART and his-torical controls indicate the need for a further detailedand systematic study of the cellular and humoral im-mune response in these patients.19,20 This study shouldinclude the analysis of more Th1- and Th2-type cyto-

kines to elucidate further the importance of the Th1-and Th2-type response in controlling JCV. Such astudy with a higher number of patients appears now tobe feasible, with a median survival of more than 46weeks in PML patients receiving HAART19,20 and theavailability of a useful JCV-authentic antigen in suffi-cient amounts.11

The study was supported by the Deutsche Forschungsgemeinschaft(We1297/3–1 and Lu397/5–1) and by a grant from the WernerOtto Stiftung to T.W. The work of M. Pickhardt is supported bythe Werner Otto Stiftung, Hamburg, Germany.

We thank Drs A. Argyrakis and C. Trebst for providing blood sam-ples and case histories of patients with PML. W.L. and T.W. con-tributed equally to this work.

References1. Astrom KE, Mancall EL, Richardson EP. Progressive multifocal

leukoencephalopathy: a hitherto unrecognized complication ofchronic lymphocytic leukemia and lymphoma. Brain 1958;81:93–111.

2. Berger JR, Concha M. Progressive multifocal leukoencephalopathy:the evolution of a disease once considered rare. J Neurovirol 1995;1:5–18.

3. Major EO, Ault GS. Progressive multifocal leukoencephalopathy:clinical and laboratory observations on a viral induced demyelinat-ing disease in the immunodeficient patient. Curr Opin Neurol1995;8:184–190.

4. Brooks BR, Walker DL. Progressive multifocal leukoencepha-lopathy. Neurol Clin 1984;2:299–313.

5. Bolton CF, Rozdilsky B. Primary progressive multifocal leu-koencephalopathy. A case report. Neurology 1971;21:72–77.

6. Berger JR, Major EO. Progressive multifocal leukoencephalop-athy. Semin Neurol 1999;19:193–200.

7. Weber T, Turner RW, Frye S, et al. Specific diagnosis of pro-gressive multifocal leukoencephalopathy by polymerase chainreaction. J Infect Dis 1994;169:1138–1141.

8. Weber T, Trebst C, Frye S, et al. Analysis of the systemic andintrathecal humoral immune response in progressive multifocalleukoencephalopathy. J Infect Dis 1997;176:250–254.

9. Horn GV, Bastian FO, Moake JL. Progressive multifocalleukoencephalopathy: failure of response to transfer factor andcytarabine. Neurology 1978;28:794–797.

10. Willoughby E, Price RW, Padgett BL, et al. Progressive multi-focal leukoencephalopathy (PML): in vitro cell-mediated im-mune responses to mitogens and JC virus. Neurology 1980;30:256–262.

11. Goldmann C, Petry H, Frye S, et al. Molecular cloning andexpression of the major structural protein VP1 of the humanpolyomavirus JC virus: formation of virus-like particles usefulfor immunological and therapeutic studies. J Virol 1999;73:4465–4469.

12. Giudici B, Vaz B, Bossolasco S, et al. Highly active antiretro-viral therapy and progressive multifocal leukoencephalopathy:effects on cerebrospinal fluid markers of JC virus replicationand immune response. Clin Infect Dis 2000;30:95–99.

13. Whiteman ML, Post MJ, Berger JR, et al. Progressive multifo-cal leukoencephalopathy in 47 HIV-seropositive patients: neu-roimaging with clinical and pathologic correlation. Radiology1993;187:233–240.

14. Weber T, Major EO. Progressive multifocal leukoencephalopathy:molecular biology, pathogenesis and clinical impact. Intervirology1997;40:98–111.

Weber et al: Immune Response in PML 641

15. Ellison GW. Progressive multifocal leukoencephalopathy (PML).I. Investigation of the immunologic status of a patient with lym-phosarcoma and PML. J Neuropathol Exp Neurol 1969;28:501–506.

16. Frye S, Trebst C, Dittmer U, et al. Efficient production of JCvirus in SVG cells and the use of purified viral antigens foranalysis of specific humoral and cellular immune response. J Vi-rol Methods 1997;63:81–92.

17. Gallia GL, Houff SA, Major EO, Khalili K. JC virus infectionof lymphocytes. J Infect Dis 1997;176:1603–1609.

18. Koralnik IJ, Boden D, Mai VX, et al. JC virus DNA load inpatients with and without progressive multifocal leukoencepha-lopathy. Neurology 1999;52:253–260.

19. Estcourt MJ, Ramshaw IA, Ramsay AJ. Cytokine response invirus infections: effects on pathogenesis, recovery, and persis-tence. Curr Opin Microbiol 1998;1:411–418.

20. Clifford DB, Yiannoutsos C, Glicksman M, et al. HAART im-proves prognosis in HIV-associated progressive multifocal leu-koencephalopathy. Neurology 1999;52:623–625.

21. Miralles P, Berenguer J, Garcia de Viedma D, et al. Treatmentof AIDS-associated progressive multifocal leukoencephalopathywith highly active antiretroviral therapy. AIDS 1998;12:2467–2472.

22. Cinque P, Casari S, Bertelli D. Progressive multifocal leukoen-cephalopathy, HIV, and highly active antiretroviral therapy.N Engl J Med 1998;339:848–849.

23. Berger JR, Levy RM, Flomenhoft D, Dobbs M. Predictive fac-tors for prolonged survival in acquired immunodeficiencysyndrome-associated progressive multifocal leukoencephalopa-thy. Ann Neurol 1998;44:341–349.

24. Gasnault J, Taoufik Y, Goujard C, et al. Prolonged survivalwithout neurological improvement in patients with AIDS-related progressive multifocal leukoencephalopathy on potentcombined antiretroviral therapy. J Neurovirol 1999;5:421–429.

25. Chaisson RE, Gallant JE, Keruly JC, Moore RD. Impact ofopportunistic disease on survival in patients with HIV infec-tion. AIDS 1998;12:29–33.

26. Dworkin MS, Wan PC, Hanson DL, Jones JL. Progressivemultifocal leukoencephalopathy: improved survival of humanimmunodeficiency virus-infected patients in the protease inhib-itor era. J Infect Dis 1999;180:621–625.

27. Berger JR, Pall L, Lanska D, Whiteman M. Progressive multi-focal leukoencephalopathy in patients with HIV infection.J Neurovirol 1998;4:59–68.

28. Szomolanyi-Tsuda E, Le QP, Garcea RL, Welsh RM. T-Cell-independent immunoglobulin G responses in vivo are elicitedby live-virus infection but not by immunization with viral pro-teins or virus-like particles. J Virol 1998;72:6665–6670.

29. Szomolanyi-Tsuda E, Welsh RM. T-cell-independent antibody-mediated clearance of polyomavirus in T-cell-deficient mice. JExp Med 1996;183:403–411.

642 Annals of Neurology Vol 49 No 5 May 2001