(PML), a fatal demyelinating disease of the central nervous system (CNS) (Berger & Concha, 1995; Hou &

Transcription

1 Journal of General Virology (2001), 82, Printed in Great Britain... Progressive multifocal leukoencephalopathy in human immunodeficiency virus type 1-infected patients: absence of correlation between JC virus neurovirulence and polymorphisms in the transcriptional control region and the major capsid protein loci Monica Sala, 1 Jean-Pierre Vartanian, 1 Pascale Kousignian, 3 Jean-François Delfraissy, 2, 3 Yassine Taoufik, 2 Simon Wain-Hobson 1 and Jacques Gasnault 2, 3 1 Unite de Re trovirologie Mole culaire, Institut Pasteur, 28 rue du Dr Roux, Paris Cedex 15, France 2 Laboratoire Virus Neurone et Immunite, Faculte de Me decine Paris Sud, 63 rue Gabriel Pe ri, Le Kremlin-Bice tre Cedex, France 3 Service de Me decine Interne, Ho pital Universitaire de Bice tre, 78 rue du Ge ne ral Leclerc, Le Kremlin-Bice tre Cedex, France Progressive multifocal leukoencephalopathy (PML) is a rapidly fatal demyelinating disease of the central nervous system related to JC polyomavirus (JCV) replication in oligodendrocytes. PML usually occurs in immunocompromised individuals, especially in the setting of AIDS. Administration of highly active anti-retroviral therapy (HAART) may improve survival prognosis in some, but not all, patients with AIDS-related PML. This observation might be explained by the outgrowth of some JCV variants of increased fitness. To evaluate this hypothesis, two subgroups of five patients with AIDS-related PML, started on HAART after PML diagnosis, were analysed. The non-responder (NR) patients died rapidly despite HAART, while responders (R) had a positive outcome and were still alive. JCV DNA was extracted from cerebrospinal fluid biopsies and two regions of the genome were analysed, the transcriptional control region (TCR) and the major capsid protein gene (VP1). Both regions show different degrees of polymorphism and are recognized as evolving independently. Sequence analysis demonstrated that (i) extensive TCR rearrangements were present in both subgroups of patients, (ii) VP1 sequence polymorphisms could be identified in the BC loop, suggesting the absence of immune selection, and (iii) no genomic marker for JCV specific neurovirulence could be identified in the TCR and VP1 loci. Introduction The JC polyomavirus (JCV) is widely distributed in the general population. The seroprevalence of JCV antibodies increases during late childhood, but primary infections appear to be asymptomatic. A latent virus infection may persist in renal tissue and JCV DNA can be detected in the urine of more than 40% of individuals older than 30 years. Against a background of immunodeficiency, JCV can replicate in oligodendrocytes, leading to progressive multifocal leukoencephalopathy (PML), a fatal demyelinating disease of the central nervous system (CNS) (Berger & Concha, 1995; Hou & Author for correspondence: Monica Sala. Fax joo pasteur.fr Major, 2000; Major & Ault, 1995). Before the AIDS epidemic, PML was a rare, opportunistic event, occurring most commonly in the setting of cellular immunodeficiency. Concurrent with the AIDS epidemic, PML prevalence has increased greatly over the past 15 years (Berger & Concha, 1995; Holman et al., 1998; Major & Ault, 1995). PML currently affects 1 4 % of AIDS patients (Berger et al., 1987; Gillespie et al., 1991; Holman et al., 1991; Major & Ault, 1995). Nowadays, the diagnosis of PML is based essentially on JCV-positive PCR from cerebrospinal fluid (CSF) (Berger & Major, 1999; Cinque et al., 1997; Weber et al., 1996). Although it is a double-stranded DNA virus, JCV manifests an extremely polymorphic portion of the transcriptional control region (TCR) that spans the origin of replication of the JCV genome through to the first ATG start codon for late gene SGM IJJ

2 M. Sala and others Fig. 1. Archetypal TCR sequence. The TCR nucleotide sequence (accession no. AF044719; Elsner & Do rries, 1998) is shown with its partition into DNA motifs A, B, C, D, E, F1 and F2 (Ault & Stoner, 1993; Elsner & Do rries, 1998). Nucleotide 1, the JCV replication origin and the first late gene ATG codon are also detailed. transcription. This includes point mutations, insertions, duplications and deletions in TCR segments that have been identified and named in the literature A E, F1 and F2 (Fig. 1) (Ault & Stoner, 1993; Elsner & Do rries, 1998). On the basis of their polymorphisms, JCV TCRs so far identified may be divided into three types. Types I (prototype Mad-1 JCV TCR) and II (prototype GS B JCV TCR) both contain repeated conserved elements that may even include the TATA box. In contrast, the so-called archetype TCR (i.e. GS K) is formed by unique sequences and a complete segment D that is located downstream of the conserved region C. TCR polymorphisms are particularly evident in brain tissue, although no correlation has been made to date with prognosis or severity. The JCV major capsid protein (VP1) is considered to be involved in virus interactions with cell receptors, haemagglutination reactions and antigenic responses. Polymorphisms within the VP1 gene are few and are generally identified at the population level rather than within an individual. On the basis of the nucleotide sequence of the VP1 upstream portion (nucleotides ; JCV strain Mad-1 taken as reference), seven major JCV genotypes have been defined (Agostini et al., 1997 c, 1998 b; Jobes et al., 1998). The JCV VP1 protein shares 77% amino acid identity with the crystallized simian virus 40 (SV40) VP1 protein [comparison made between SV40 VP1 (Liddingtonet al., 1991) and JCV strain Taiwan-3, accession no. U61771 (Chang et al., 1996)]. By analogy with the SV40 VP1 protein, three major structural loops can be identified within JCV VP1, the BC loop (amino acids 48 81; Mad-1 genome as reference), the DE loop (amino acids ) and the HI loop (amino acids ). Two deletions of eight and nine amino acids in the DE loop (amino acids and , respectively) have been described as dominant JCV forms in two AIDS patients developing PML (Stoner & Ryschkewitsch, 1995). In both patients, each deleted JCV clone showed better fitness and a replication advantage in comparison with minor, undeleted forms. Polymorphisms in the TCR and VP1 loci have been demonstrated to evolve independently (Agostini et al., 1997b). Before the introduction of highly active anti-retroviral therapy (HAART) directed against human immunodeficiency virus (HIV), the median survival time of AIDS patients after the clinical onset of PML was about 4 months (Berger & Concha, 1995; Fong & Toma, 1995; Gasnault et al., 1999; Gillespie et al., 1991). In recent series studies (Albrecht et al., 1998; Cinque et al., 1998; Clifford et al., 1999; Gasnault et al., 1999; Miralles et al., 1998) and cohort studies (Dworkin et al., 1999; Tassie et al., 1999), HAART has led to a significant improvement in median patient survival time to more than 1 year. However, this encouraging improvement applies only to about one case in two (Taoufik et al., 2000). At PML diagnosis, low CD4+ T lymphocyte count (Albrecht et al., 1998; Clifford et al., 1999; Fong & Toma, 1995; Gasnault et al., 1999) and high JCV load in CSF (Taoufik et al., 1998, 2000) have been correlated with short survival time, despite sustained control of HIV replication under HAART. With respect to outcome following the onset of HAART, it could be hypothesized that, in addition to acquired immunodeficiency and host specificities, virus features possibly related to specific JCV polymorphisms could influence the course of PML as a result of changes in JCV fitness. Accordingly, the TCR and the VP1 loci were analysed from JCV obtained from CSF of ten patients with AIDS-related PML. Methods DNA from 100 µl CSF from the patients described in Table 1 was precipitated with ethanol after adding 25 µl salmon sperm DNA (100 µg ml) and then resuspended in 50 µl pure water. From each sample, JAA

3 Table 1. Patient clinical parameters and outcomes Abbreviations: D, deceased; A, alive; CSF, cerebrospinal fluid; AZT, zidovudine; 3TC, lamivudine; D4T, stavudine; DDI, didanosine; ABC, abacavir; NVP, nevirapine; IDV, indinavir; RTV, ritonavir; NFV, nelfinavir. CD4+ cell HIV RNA (copies/ml) Survival count CSF JCV DNA Patient Clinical symptoms (months)* Outcome (/µl) Plasma CSF (copies/ml) Anti-HIV regimen JAB P1 Frontal lobe syndrome, left hemiparesia, swallowing 3 2 D AZT 3TC IDV disorders P5 Diplopia, left cerebellar ataxia, dysarthria, swallowing 3 4 D TC ABCV RTV IDV disorders P8 Subcortical blindness, right hemiparesia, aphasia 3 5 D D4T DDI IDV P10 Frontal lobe syndrome, right hemiparesia, aphasia 4 1 D AZT 3TC IDV P12 Frontal lobe syndrome, right hemiplegia, swallowing 4 5 D AZT 3TC NVP NFV disorders P2 Amnesia, Balint s syndrome, right hemiparesia, aphasia 17 5 A AZT 3TC IDV P4 Frontal lobe syndrome, right facial paresis, aphasia 24 7 A TC D4T P7 Frontal lobe syndrome, aphasia, swallowing disorders 14 8 A AZT 3TC RTV IDV P9 Bilateral cerebellar ataxia, dysarthria 37 5 A AZT 3TC IDV P11 Right homonymous hemianopia, dysgraphia, right cerebellar ataxia 15 8 A AZT 3TC NFV * Following clinical onset of PML. At PML diagnosis. PML outcome and JC virus polymorphisms

4 M. Sala and others Table 2. TCR sequences and motif organization The various rearranged motifs indicated are: B, archetypal motif B with 5 bp deleted downstream of the 5 -GGGAGG sequence; B, archetypal B motif lacking 6 bp at the 5 end; C, C motif with sequence 5 -GCCAAGCATGAGCTC duplicated; C, archetypal C motif lacking 19 (P8-2-48) or 23 (P7-2-23) bp at the 3 end; C, archetypal C motif lacking 16 bp at the 5 end; C*, highly truncated C motif represented by the sequence 5 -ATACCT; D,12 bp (5 -TCCTTAATCACA) 3 -truncated form of the 66 bp archetypal D motif; D, 9 bp (5 -TTAATCACA) 3 -truncated form of the archetypal D motif; E, archetypal motif E lacking 6 bp at the 3 end; E, archetypal E motif lacking 6 bp at the 5 end; F1*, 25 bp (5 -GAACATGTTTTGCGAGCCAGAGCTG) 5 -truncated form of the 44 bp archetypal F1 motif; F1, 9 bp deleted form of the 44 bp archetypal F1 motif (deletion starts 4 bp downstream of the 5 end of F1). Frequency Sequence Motif organization n % Non-responders P1-1-1 A B C B C D E B C D E F1 F P1-1-9 A B C D E F1 F2 1 8 P1-3-1 A B C B C D E F1 F P5-2-7 A B C D E F1 F P A B C B C D E F1 F P A B C E F1 F P A B C E B C E F1 F P A B C D E F1 F2 1 5 P A B C B C D E F1* F P A B C E B C E F1 F P A B C E F1 F P A B C E B C E B C E F1 F2 1 4 Responders P A B C E B C E F1 F P A B C D E F1 F P A B C E F1 F2 1 6 P A B C C D E F1 F2 1 6 P A B C E B C E F1 F P A C E B C E B C E F1 F P A B C E F1 F P A B C E B C E F1 F P A B C E C* E F1 F P A B C E B C E F1 F P A B C E F1 F P A B C D E F1 F P A B C B C E F1 F2 1 8 P A B C D E C D E F1 F P A B C D E F1 F The number of clones from that patient having the TCR rearrangement detailed (n) and the frequency of that sequence amongst the clones from each patient (%) are shown. Major TCR sequence for each patient. 20 µl was used for a nested PCR to amplify the JCV TCR and VP1 regions. The outer primers for the TCR region were R5 E (5 - GGCGGAATTCTGGATTCCTCCCTA-3 ) and R3 E (5 -GGCGCTG- CAGACAGAAGCCTTACGTGACAG-3 ) and the inner primers were R5 I (5 -GGCGGAATTCCTTCTGAGTAAGCTTGGAGGCGG-3 ) and R3 I (5 -GGCGCTGCAGGGCGAAGAACCATGGCCAGCTGG- 3 ). The outer primers for the VP1 region were VP15 E (5 -GGCGG- GATCCATGGCCCCAACAAAAAGAAAAGG-3 ) and VP13 E (5 - GGCGAAGCTTGCTGGTTATACTTTATTAAAATGTACTG-3 ) and the inner primers were VP15 I (5 -GGCGGGATCCGAAAGGAAGG- ACCCCGTGCAAG-3 ) and VP13 I(5 -GGCGAAGCTTCCAACAGA- AAAAAAATGAAAGCTGGTG-3 ). Optimized PCR conditions were 20 mm Tris HCl (ph 8 8), 10 mm KCl, 10 mm (NH )SO, 2 mm MgSO,0 1% Triton X-100, 100 µg ml nuclease-free BSA and 2 5 U cloned Pfu DNA polymerase in a final volume of 100 µl. PCR tubes were heated at 95 C for 5 min and then subjected to 40 cycles of denaturation (95 C for 30 s), annealing (60 C for 30 s) and extension (74 C for 5 min). In the last PCR cycle, the extension step lasted for 10 min. PCR products were precipitated with ethanol and resuspended in 70 mm Tris HCl, ph 7 6, 10 mm MgCl and 5 mm DTT and incubated at 37 C for 1 h with 20 U T4 polynucleotide kinase (BioLabs). After inactivation of the enzyme at 65 C for 20 min, phosphorylated PCR products were cloned blunt-ended into SmaI-digested M13mp18 plasmid. Plaques were screened for JCV inserts by using the following probes: for the TCR region, RS1 (5 -AGTGGAAAGCAGCCAAGGGAACATG-3 ) and RS2 (5 -ATGGCTGCCAGCCAAGCATGAGCTCA-3 ), and for the JAC

5 PML outcome and JC virus polymorphisms VP1 region, VP1S (5 -ATTACAGAGGTAGAATGCTTTTTA-3 ) and SVP1 (5 -CTTATAAGAGGAGGAGTAG-3 ). Between 10 and 25 clones of the TCR and VP1 regions of each patient were collected and sequenced with the DYEnamic direct cycle sequencing kit, the 21 M13 forward primer (Amersham) and an Applied Biosystems 373A sequencer. VP1 sequences were aligned, a distance matrix was calculated by using Clustal W version 1.7 and phylogenetic trees were obtained by using the Neighbor program of the PHYLIP package version 3.5. All sequences obtained in this study as well as their alignments are available through anonymous login at ftp.pasteur.fr pub retromol Sala00. Results Patients and their clinical backgrounds This study was performed on CSF samples from patients with AIDS-related PML. All patients were documented for HIV infection. PML diagnosis was based on the following criteria: (i) focal brain disease with subacute progression, (ii) white-matter lesions on magnetic resonance imaging consistent with PML, (iii) detection of JCV in CSF by PCR and (iv) there being no other likely aetiology. Before onset of PML, patients were either untreated or under nucleoside analogue mono- or bi-therapy. Following PML diagnosis, all patients were started on HAART, including a protease inhibitor. With respect to clinical outcome following the onset of HAART, two representative groups of five patients each were selected from 39 HIV-1-infected patients who were diagnosed with PML between May 1996 and September 1999 and from whom a sufficient CSF sample was drawn at PML diagnosis and stored at 80 C (Table 1) (Taoufik et al., 2000). Despite HAART, the five non-responder patients (NR subgroup) died within a mean time of 3 7 months after PML diagnosis. The five responder patients (R subgroup) were still alive at the end of 1999, presenting a mean survival time of 22 1 months. For both groups, the CSF samples were drawn at a median time of 1 2 months after clinical onset of PML. JCV load in the CSF was measured by quantitative PCR, as described previously (Taoufik et al., 1998). CD4+ T cell counts and CSF and plasma HIV-1 loads were determined at PML diagnosis. Trends towards higher JCV load (means 6 5 log JCV DNA copies ml compared with 4 5 log JCV DNA copies ml in R patients) and lower CD4+ cell counts (means 94 cells µl vs 179 cells µl in R patients) were observed in NR patients (Table 1). The marked difference in survival between R and NR patients was not related to non-adherence to or failure of HAART. In fact, at 3 months after onset of combined therapy, similar decreases in the plasma HIV load were observed in the two patient subgroups (data not shown). Transcriptional control region (TCR) Regardless of patient group, the ten samples showed sequence polymorphisms in the TCR, with between two and four different sequence prototypes per sample (Table 2; see also ftp.pasteur.fr pub retromol Sala00). Within each patient, the A, B, C, E, F1 and F2 motifs (Ault & Stoner, 1993; Elsner &Do rries, 1998) were generally highly conserved, with most of the length polymorphisms occurring between the 5 -C and 3 -E motifs, where insertions and deletions were generally confined. By virtue of the presence of a single TATA box, all were type II TCRs. None of the sequences showed an archetypal TCR: all seven motif sequences had deletions in the D segment (D and D ; Table 2). Given the size of the sample (173 TCR sequences) and the titres of JCV DNA (Table 1), the clonal frequencies reflect the real frequencies in the CSF Table 3. Specific nucleotides corresponding to VP1 nucleotide positions that allow JCV genotype definition Underlined residues indicate point mutations from the genotype prototype. VP1 nucleotide residue No. of JCV Mutation from Sequence clones genotype JCV type Non-responders P1 A C G T T C T G G C A 7 T4 A G P5 A C G T T C T G G C A 23 T4 A G P8 A C G T T G T T G G G 21 T1B P10 A C G T T G T T G G G 20 T1B P12 A C G T T G T G A G G 20 T1A Responders P2 A C G T T G T T G G A 22 T1B G A P4 A C G T T G T T G G G 15 T1B P7 A A T T T C C T G G A 20 T2B P9 A C G T T G T T G G G 19 T1B P11-72 A C G T T G T T G G G 9 T1B P11-73 A C G T T G T T G G G 16 T1B P11-74 A C G T T G T T G G G 9 T1B JAD

6 M. Sala and others Table 4. VP1 amino acid residue variations among JCV strains The BC loop, β-d, α-b and DE loop are JCV structural elements defined by amino acid sequence similarity to SV40 VP1 (Chang et al., 1996). VP1 amino acid residue BC loop β-d α-b DE loop Sequence Non-responders P1 F S N T T A P5 F S N T T A P8 L S S T S G P10 L S S T S G P12 L S S T S G Responders P2 F P S T S A P4 L P S T S G P7 L S N T T A P9 L S S N S A P11-72 F P S T S G P11-73 L P S T S G P11-74 F S S T S G reasonably well. However, there was no defining JCV molecular trait that distinguished the two groups of patients, nor any distinction in the relative frequencies of major and minor TCR forms. Major capsid protein (VP1) In the present study, 201 VP1 sequences (about 20 sequences per patient) were analysed (Tables 3 and 4; see also ftp.pasteur.fr pub retromol Sala00). The region amplified from VP1 covers all the nucleotide positions used to distinguish between JCV major genotypes (types 1 7) and JCV subtypes of types 1 and 2 (Agostini et al., 1997c, 1998b; Jobes et al., 1998). Moreover, this region of VP1 encompasses two major protein structural loops that can be defined by amino acid sequence similarity to SV40 VP1, the BC and DE loops (Chang et al., 1996; Liddington et al., 1991). Most patients were infected by JCV strains of the T1B subtype, but T1A, T2B and T4 subtypes and types were also represented (Table 3) (Agostini et al., 1996b, 1998b; Jobes et al., 1998). At one determinant position for JCV typing, VP1 sequences from patients P1, P5 and P2 showed a nucleotide variation from the closest typological prototype (Table 3). Nucleotide and amino acid substitutions within VP1 sequences failed to distinguish the two subgroups of patients (Tables 3 and 4). Within each individual, VP1 sequences were extremely conserved. The only exception to this was patient P11. Three VP1 nucleotide and amino acid sequence prototypes could be identified in the sample from this patient, corresponding to clones P11-72, P11-73 and P11-74 (Table 4). At nucleotide positions 1631 and 1649 (numbering according to the complete Mad-1 genome sequence), P11-72 had nucleotides T and C, P11-73 had C and C and P11-74 had T and T. These point variations led to different amino acids being encoded at positions 55 and 61 (Table 4). Both amino acid positions are in the VP1 BC loop, which is considered to be one of the major sites involved in JCV antibody recognition and receptor interaction (Chang et al., 1996). Discussion The factors that influence JCV neurotropism and neurovirulence remain unknown. It has been suggested that it may be possible to identify prognosis markers for PML progression in polymorphic regions of the JCV genome. To this end, JCV DNA amplified from CSF samples obtained from patients with AIDS-related PML were cloned and sequenced. The study focused on the TCR and the VP1 polymorphic regions, which are respectively fundamental in regulating virus transcription and in intervening in virus host interactions. Transcriptional control region (TCR) Within the TCR, the promoter enhancer area between the origin of replication (ori) and the start codon of the agnoprotein gene can be highly rearranged. Rearranged TCR forms are considered to be derived from the archetypal form through multiple recombination events (Yogo et al., 1990). Efforts to describe the distribution of rearranged forms in different body compartments and to define a correlation between rearranged forms and PML outcome have yielded controversial results (Elsner & Do rries, 1998). JCV sequences from AIDS patients with and without PML showed rearranged TCR forms and rearranged forms could also be detected in kidney, lung, spleen and urine samples (Caldarelli-Stefano et al., 1999; Elsner & Do rries, 1998). These last data indicate that the emergence of rearranged forms of the JCV TCR is not necessarily correlated with JCV colonization of CNS and PML outcome. Moreover, the present study does not support the hypothesis that particular TCR variants actually represent additional risk factors for the occurrence of PML in immunocompromised individuals. Analysis of TCR sequences from ten patients with AIDS-related PML who were on HAART, including five with a rapid fatal outcome and five with no progression in PML course, failed to reveal any TCR form that was specific to each subgroup of patients, despite extensive polymorphisms. Independent of their response to HAART, each patient showed a heterogeneous distribution of TCR variants in CSF. In most patients, one form predominated over minor TCR forms that seemed to be derived from the major form. Only in patients P2 and P9 could two different TCR rearrangements compete with apparently comparable fitness. JAE

7 PML outcome and JC virus polymorphisms Three TCR sequences deserve special comment. P and P showed deletions in the F1 segment that eliminated the pseudo-nf-1 site (5 -TGGAAAGCAGCCA-3 ) (see legend to Table 2). It can be deduced that these deletions do not affect JCV replication in the CNS, since sequences with the pseudo-nf-1 site deleted were detected as the major virus forms in patients P10 and P7 (Table 2). In P4-5-34, the B motif next to motif A was deleted entirely and it was present twice downstream between the E and C segments. This deletion brings the A and C motifs close one to the other and results in the reconstitution of the Bpenta region (5 -AGGGAAGGGA- 3 ) (Raj & Khalili, 1995). In all other TCR sequences, this region is split into two parts, the first (5 -AGGGAAGG-3 ) in the A motif and the second (5 -GA-3 ) in the C motif, separated by the B segment (Fig. 1). The Bpenta region has been shown to be important for DNA replication and for gene translation, functioning as an activator of the JCV early promoter and as a repressor of the JCV late promoter (Raj & Khalili, 1995). The P sequence type was identified only in one patient and even then as a minor TCR form (11% of P4 clones). Therefore, it would appear that the existence of more than one copy of the AGGGA pentamer (already present in the A motif) is far from essential for massive virus replication. Moreover, data show that, apart from the minor P sequence, all other major or minor sequences presented the A B C motif organization at the 5 -end of the analysed TCR. Hence, this seems to result in better fitness for the virus, at least in the CNS. Major capsid protein (VP1) The analysis of 201 VP1 sequences failed to show any marker for neurovirulence at this locus. Sequences were highly conserved, presenting no deletions (Stoner & Ryschkewitsch, 1995) or other major sequence modifications (insertion, duplication). Few differing nucleic acid or amino acid residues were identified, and they are described in Tables 3 and 4. At each variable position, the amino acid choice appeared very narrow, always being between a maximum of two residues. JCV type 2B strains have been suggested to be associated with a higher risk of PML (Agostini et al., 1997 a, 1998 a). The present data show that no correlation can be made between virus genotype and JCV specific neurovirulence (Table 3). The VP1 sequences obtained in the study were aligned and a distance matrix was obtained to verify whether the NR and R sequences grouped separately in a phylogenetic tree. Of distances ranging from 0 3 to2 1%, the intra- and intergroup variation was not significantly different (data not shown). Within each CSF sample, VP1 nucleotide sequences were extremely conserved. The only exception was the P11 sequence set, where three VP1 sequences co-existed. By comparison with the other samples, these sequences are as divergent as those from distinct individuals (Table 4). These sequences, P11-72, P11-73 and P11-74, represented 26, 47 and 26% of the P11 sample. This percentage determination was highly significant, as it was determined on a large sample (34 sequences) and by a reliable methodology for DNA amplification and clone screening (Meyerhans et al., 1990). The Pfu DNA polymerase used for DNA amplification possesses a 3 5 exonuclease proof-reading activity that enables the polymerase to correct nucleotide misincorporations. It is unlikely that one of the three P11 clones could have been the origin of the other two in vivo, since two nucleotide point mutations would be required. This would imply an intrapatient non-synonymous nucleotide substitution frequency (K A ) in the VP1 region analysed of about 5 10 bp. This corresponds to the estimated K A value from VP1 gene comparisons among JCV types spread across geographically distinct human populations (Hatwell & Sharp, 2000), while the intrapatient VP1 substitution frequency is assumed to be far smaller. Given the presence of residues F-55 and P-61 (Table 4), sequence P11-72 looks to be a recombinant clone obtained from sequences P11-73 and P This means that susceptible cells may be super-infected or co-infected by a number of JCV virions. Recombination clearly may occur in all cases: it is simply that variation is necessary to reveal it. When polymorphisms in the TCR region arise by recombination and or deletion insertion events, the only unknown is whether they arose by intra- or intermolecular recombination. The above finding from the VP1 and TCR data sets suggests that both probably operate. The three independent JCV clones from patient P11 are of the same subtype, T1B (Table 3). Patients co-infected with more than one JCV genotype in a single tissue have been described in the literature (Agostini et al., 1996a, b; Ault & Stoner, 1992). The T4 genotype itself, characterized by the replacement of a region of the T1 sequence by a fragment of the African JCV T3 sequence, provides direct evidence that coinfection of the same cell and consequent recombination between different virus types do occur during JCV replication. Here, it is shown that co-infection in the CNS is possible, even between JCV clones of the same subtype. The amino acid positions at which the three P11 clones vary are all in the BC loop (Table 4). This loop is considered to be an antigenic region for polyomavirus VP1 proteins (Chang et al., 1996). The present data suggest either that these amino acid positions are not involved in epitope recognition or that immune responses to JCV infection are irrelevant as far as JCV establishment and replication in the brain are concerned. In conclusion, it is demonstrated here that no specific genetic marker in the TCR or VP1 loci can be correlated with JCV neurovirulence. Despite extensive polymorphisms, sequence analysis failed to reveal any significant relationship between TCR or VP1 polymorphisms and PML outcome. However, it cannot be ruled out that the principal locus affecting JCV neurovirulence lies elsewhere in the JCV genome. Alternatively, in the era of HAART, the different outcome of PML might be related to the dynamics of anti-jcv immune reconstitution, determined by host specificities, and or to the severity of acquired immunodeficiency. JAF

8 M. Sala and others We are grateful to Michel Henry and Mufide Kahraman for technical support. This work was supported by the Institut Pasteur, l Agence Nationale de Recherche sur le SIDA (ANRS), SIDACTION, INSERM and Universite Paris Sud. References Agostini, H. T., Ryschkewitsch, C. F., Singer, E. J. & Stoner, G. L. (1996a). Co-infection with two JC virus genotypes in brain, cerebrospinal fluid or urinary tract detected by direct cycle sequencing of PCR products. Journal of Neurovirology 2, Agostini, H. T., Ryschkewitsch, C. F. & Stoner, G. L. (1996b). Genotype profile of human polyomavirus JC excreted in urine of immunocompetent individuals. Journal of Clinical Microbiology 34, Agostini, H. T., Ryschkewitsch, C. F., Mory, R., Singer, E. J. & Stoner, G. L. (1997a). JC virus (JCV) genotypes in brain tissue from patients with progressive multifocal leukoencephalopathy (PML) and in urine from controls without PML: increased frequency of JCV type 2 in PML. Journal of Infectious Diseases 176, 1 8. Agostini, H. T., Ryschkewitsch, C. F., Singer, E. J. & Stoner, G. L. (1997b). JC virus regulatory region rearrangements and genotypes in progressive multifocal leukoencephalopathy: two independent aspects of virus variation. Journal of General Virology 78, Agostini, H. T., Yanagihara, R., Davis, V., Ryschkewitsch, C. F. & Stoner, G. L. (1997c). Asian genotypes of JC virus in Native Americans and in a Pacific Island population: markers of viral evolution and human migration. Proceedings of the National Academy of Sciences, USA 94, Agostini, H. T., Ryschkewitsch, C. F., Singer, E. J., Baumhefner, R. W. & Stoner, G. L. (1998a). JC virus type 2B is found more frequently in brain tissue of progressive multifocal leukoencephalopathy patients than in urine from controls. Journal of Human Virology 1, Agostini, H. T., Shishido-Hara, Y., Baumhefner, R. W., Singer, E. J., Ryschkewitsch, C. F. & Stoner, G. L. (1998b). JC virus Type 2: definition of subtypes based on DNA sequence analysis of ten complete genomes. Journal of General Virology 79, Albrecht, H., Hoffmann, C., Degen, O., Stoehr, A., Plettenberg, A., Mertenskotter, T., Eggers, C. & Stellbrink, H. J. (1998). Highly active antiretroviral therapy significantly improves the prognosis of patients with HIV-associated progressive multifocal leukoencephalopathy. AIDS 12, Ault, G. S. & Stoner, G. L. (1992). Two major types of JC virus defined in progressive multifocal leukoencephalopathy brain by early and late coding region DNA sequences. Journal of General Virology 73, Ault, G. S. & Stoner, G. L. (1993). Human polyomavirus JC promoter enhancer rearrangement patterns from progressive multifocal leukoencephalopathy brain are unique derivatives of a single archetypal structure. Journal of General Virology 74, Berger, J. R. & Concha, M. (1995). Progressive multifocal leukoencephalopathy: the evolution of a disease once considered rare. Journal of Neurovirology 1, Berger, J. R. & Major, E. O. (1999). Progressive multifocal leukoencephalopathy. Seminars in Neurology 19, Berger, J. R., Kaszovitz, B., Post, M. J. & Dickinson, G. (1987). Progressive multifocal leukoencephalopathy associated with human immunodeficiency virus infection. A review of the literature with a report of sixteen cases. Annals of Internal Medicine 107, Caldarelli-Stefano, R., Vago, L., Omodeo-Zorini, E., Mediati, M., Losciale, L., Nebuloni, M., Costanzi, G. & Ferrante, P. (1999). Detection and typing of JC virus in autopsy brains and extraneural organs of AIDS patients and non-immunocompromised individuals. Journal of Neurovirology 5, Chang, D., Liou, Z.-M., Ou, W.-C., Wang, K.-Z., Wang, M., Fung, C.-Y. & Tsai, R.-T. (1996). Production of the antigen and the antibody of the JC virus major capsid protein VP1. Journal of Virological Methods 59, Cinque, P., Scarpellini, P., Vago, L., Linde, A. & Lazzarin, A. (1997). Diagnosis of central nervous system complications in HIV-infected patients: cerebrospinal fluid analysis by the polymerase chain reaction. AIDS 11, Cinque, P., Casari, S. & Bertelli, D. (1998). Progressive multifocal leukoencephalopathy, HIV, and highly active antiretroviral therapy. New England Journal of Medicine 339, Clifford, D. B., Yiannoutsos, C., Glicksman, M., Simpson, D. M., Singer, E. J., Piliero, P. J., Marra, C. M., Francis, G. S., McArthur, J. C., Tyler, K. L., Tselis, A. C. & Hyslop, N. E. (1999). HAART improves prognosis in HIV-associated progressive multifocal leukoencephalopathy. Neurology 52, Dworkin, M. S., Wan, P.-C., Hanson, D. I. & Jones, J. L. (1999). Progressive multifocal leukoencephalopathy: improved survival of human immunodeficiency virus-infected patients in the protease inhibitor era. Adult and Adolescent Spectrum of HIV Disease Project. Journal of Infectious Diseases 180, Elsner, C. & Do rries, K. (1998). Human polyomavirus JC control region variants in persistently infected CNS and kidney tissue. Journal of General Virology 79, Fong, I. W. & Toma, E. (1995). The natural history of progressive multifocal leukoencephalopathy in patients with AIDS. Canadian PML Study Group. Clinical Infectious Diseases 20, Gasnault, J., Taoufik, Y., Goujard, C., Kousignian, P., Abbed, K., Boue, F., Dussaix, E. & Delfraissy, J. F. (1999). Prolonged survival without neurological improvement in patients with AIDS-related progressive multifocal leukoencephalopathy on potent combined antiretroviral therapy. Journal of Neurovirology 5, Gillespie, S. M., Chang, Y., Lemp, G., Arthur, R., Buchbinder, S., Steimle, A., Baumgartner, J., Rando, T., Neal, D., Rutherford, G. and others (1991). Progressive multifocal leukoencephalopathy in persons infected with human immunodeficiency virus, San Francisco, Annals of Neurology 30, Hatwell, J. N. & Sharp, P. M. (2000). Evolution of human polyomavirus JC. Journal of General Virology 81, Holman, R. C., Janssen, R. S., Buehler, J. W., Zelasky, M. T. & Hooper, W. C. (1991). Epidemiology of progressive multifocal leukoencephalopathy in the United States: analysis of national mortality and AIDS surveillance data. Neurology 41, Holman, R. C., To ro k, T. J., Belay, E. D., Janssen, R. S. & Schonberger, L. B. (1998). Progressive multifocal leukoencephalopathy in the United States, : increased mortality associated with HIV infection. Neuroepidemiology 17, Hou, J. & Major, E. O. (2000). Progressive multifocal leukoencephalopathy: JC virus induced demyelination in the immune compromised host. Journal of Neurovirology 6 (Suppl. 2), S98 S100. Jobes, D. V., Chima, S. C., Ryschkewitsch, C. F. & Stoner, G. L. (1998). Phylogenetic analysis of 22 complete genomes of the human polyomavirus JC virus. Journal of General Virology 79, Liddington, R. C., Yan, Y., Moulai, J., Sahli, R., Benjamin, T. L. & Harrison, S. C. (1991). Structure of simian virus 40 at 3 8-A resolution. Nature 354, JAG

9 PML outcome and JC virus polymorphisms Major, E. O. & Ault, G. S. (1995). Progressive multifocal leukoencephalopathy: clinical and laboratory observations on a viral induced demyelinating disease in the immunodeficient patient. Current Opinion in Neurology 8, Meyerhans, A., Vartanian, J.-P. & Wain-Hobson, S. (1990). DNA recombination during PCR. Nucleic Acids Research 18, Miralles, P., Berenguer, J., Garcia de Viedma, D., Padilla, B., Cosin, J., Lopez-Bernaldo de Quiros, J. C., Munoz, L., Moreno, S. & Bouza, E. (1998). Treatment of AIDS-associated progressive multifocal leukoencephalopathy with highly active antiviral therapy. AIDS 12, Raj, G. V. & Khalili, K. (1995). Transcriptional regulation: lessons from the human neurotropic polyomavirus, JCV. Virology 213, Stoner, G. L. & Ryschkewitsch, C. F. (1995). Capsid protein VP1 deletions in JC virus from two AIDS patients with progressive multifocal leukoencephalopathy. Journal of Neurovirology 1, Taoufik, Y., Gasnault, J., Karaterki, A., Pierre Ferey, M., Marchadier, E., Goujard, C., Lannuzel, A., Delfraissy, J. F. & Dussaix, E. (1998). Prognostic value of JC virus load in cerebrospinal fluid of patients with progressive multifocal leukoencephalopathy. Journal of Infectious Diseases 178, Taoufik, Y., Delfraissy, J.-F. & Gasnault, J. (2000). Highly active antiretroviral therapy does not improve survival of patients with high JC virus load in the cerebrospinal fluid at progressive multifocal leukoencephalopathy diagnosis. AIDS 14, Tassie, J. M., Gasnault, J., Bentata, M., Deloumeaux, J., Boue, F., Billaud, E. & Costagliola, D. (1999). Survival improvement of AIDSrelated progressive multifocal leukoencephalopathy in the era of protease inhibitors. Clinical Epidemiology Group. French Hospital Database on HIV. AIDS 13, Weber, T., Frye, S., Bodemer, M., Otto, M. & Luke, W. (1996). Clinical implications of nucleic acid amplification methods for the diagnosis of viral infections of the nervous system. Journal of Neurovirology 2, Yogo, Y., Kitamura, T., Sugimoto, C., Ueki, T., Aso, Y., Hara, K. & Taguchi, F. (1990). Isolation of a possible archetypal JC virus DNA sequence from nonimmunocompromised individuals. Journal of Virology 64, Received 19 September 2000; Accepted 4 December 2000 Published ahead of print (19 December 2000) in JGV Direct as DOI /vir JAH