OUP user menu

A prospective study demonstrates an association between JC virus-specific cytotoxic T lymphocytes and the early control of progressive multifocal leukoencephalopathy

Renaud A. Du Pasquier, Marcelo J. Kuroda, Yue Zheng, Jims Jean-Jacques, Norman L. Letvin, Igor J. Koralnik
DOI: http://dx.doi.org/10.1093/brain/awh215 1970-1978 First published online: 23 June 2004

Summary

Progressive multifocal leukoencephalopathy (PML) is a fatal demyelinating disease of the CNS of immunosuppressed individuals caused by the polyomavirus JC (JCV). In previous studies, we showed that JCV-specific cytotoxic T lymphocytes (JCV-specific CTL) were associated with a favourable outcome in patients with PML. However, these CTL had been assessed in PML survivors more than 1 year after the onset of disease and we could not determine whether this immune response was only a surrogate marker for a general recovery of the patient's immune system or a causal factor in the patient's neurological improvement. In this study, we assessed the relationship between JCV-specific CTL detected early in the course of PML and the subsequent course of disease activity. We enrolled 26 patients with possible or proven PML, including 21 HIV+ patients, less than 10 months after the onset of their neurological symptoms (3.7 ± 2.5 months, median ± interquartile range). JCV-specific CTL were detected by either 51Cr release or tetramer staining assay. Patients were then followed prospectively and the clinical course of PML was determined. At the time of their first immune evaluation, we found that 15 patients had detectable JCV-specific CTL. HIV+ patients with JCV-specific CTL had a higher CD4+ T-cell count (215 ± 103/µl) and a lower HIV viral load (144 ± 431 copies/ml) than those without JCV-specific CTL (32 ± 59/µl, P = 0.004 and 43 100 ± 54 778 copies/ml, P = 0.01). Thirteen of these 15 patients with JCV-specific CTL developed clinically quiescent PML, while only two out of 11 without detectable CTL controlled their neurological disease. Therefore, the early detection of JCV-specific CTL had an 87% predictive value for subsequent control of PML, while the absence of such CTL had an 82% predictive value for subsequent active PML (P = 0.0009). Fifteen patients were evaluated less than 4 months after the onset of PML (1.9 ± 1.3 months). Of nine patients with JCV-specific CTL, seven (78%) demonstrated subsequent control of disease, whereas six out of six (100%) without JCV-specific CTL developed progressive PML (P = 0.007). Two to ten CTL assays were performed on PBMC of 11 patients. Of these patients, one had an increase in JCV-specific CTL preceding a significant clinical improvement. In another patient with otherwise stable immune parameters, a decline in JCV-specific CTL preceded an exacerbation of PML. We conclude that JCV-specific CTL can be detected early in PML and can predict control of this disease. Fluctuations of JCV-specific CTL in the blood are associated with variation in disease manifestations. These results indicate that JCV-specific CTL are associated with the control of PML.

  • cellular immune response
  • demyelinating disorders
  • HIV
  • JC virus
  • progressive multifocal leukoencephalopathy
  • CTL = cytotoxic T lymphocytes
  • HAART = highly active antiretroviral therapy
  • HIV Gagp77-specific CTL = HLA-A* 0201-restricted CTL epitope on the Gag protein of HIV
  • IQR = interquartile range
  • JCV = JC virus
  • JCV VPI-sp CTL = HLA-A* 0201-restricted CTL epitopes of the JCV virus
  • JCV-specific CTL = JCV-specific cytotoxic lymphocytes
  • PBMC = peripheral blood mononuclear cells
  • PML = progressive multifocal leukoencephalopathy

Introduction

The polyomavirus JC (JCV) infects more than 85% of adult humans (Weber et al., 1997). After primary infection, JCV remains quiescent in the kidneys and lymphoid organs. In healthy individuals, this virus can replicate in tubular kidney cells and is excreted in the urine, without causing any disease. However, in the context of severe immunosuppression, in organ transplant recipients, cancer patients or people with AIDS, JCV can spread to the CNS and cause a severe, often fatal, demyelinating disease called progressive multifocal leukoencephalopathy (PML).

Despite the availability of highly active antiretroviral therapy (HAART), which can markedly improve immune function of HIV+ patients, the mortality associated with PML is approximately 50% within 1 year of diagnosis (Clifford et al., 1999). Therefore, determination of reliable prognostic factors of disease evolution is of crucial importance for patient management. Berger and colleagues showed that in patients who had PML heralding AIDS, those with a CD4+ T-cell count greater than 300/µl and contrast-enhancing lesions on MRI had prolonged survival (Berger et al., 1998). These findings suggest that a vigorous immune response against JCV may confer protection against disease progression.

In previous studies, we have shown that the presence of JCV-specific cytotoxic T lymphocytes (JCV-specific CTL) was associated with a favourable outcome in patients with PML (Du Pasquier et al., 2001; Koralnik et al., 2001). We have characterized two HLA-A*0201-restricted CTL epitopes of the major capsid protein of JCV (JCV VP1-sp CTL), VP1p100 (Koralnik et al., 2002) and VP1p36 (Du Pasquier et al., 2003). Moreover, JCV-specific CTL were detected in 10 of 11 PML survivors (91%) and one of 11 progressors (9%, P = 0.0003) (Du Pasquier et al., 2003). Since almost all PML survivors enrolled in these studies were clinically stable or improving neurologically for more than 1 year at the time of the immune evaluation, we could not rule out the possibility that detectable JCV-specific CTL were merely surrogate markers reflecting a general recovery of the immune function of these patients rather than a causal factor in their neurological improvement.

To determine if JCV-specific CTL arise early in the course of PML and to ascertain their role in the containment of disease progression, we evaluated the cellular immune response of 26 PML patients shortly after disease onset. These patients were then followed prospectively and 11 had more than one CTL assay performed. We found that the appearance of JCV-specific CTL was an early event in PML and a reliable prognostic marker of subsequent disease control. We also showed in one patient that expansion of his cell population preceded significant clinical improvement, and a fall in the number of these cells heralded clinical deterioration in another.

Methods

Selection of study subjects

To determine the role played by JCV VP1-sp CTL early in the course of PML, we enrolled 26 patients with possible or proven PML, less than 10 months after disease onset. All patients gave written consent according to our institution's review board (IRB) guidelines. The patients were enrolled after May 1, 2000 and were followed clinically up to the end-point of the study, which was either September 30, 2003 or the death of the patient. Eleven patients had more than one CTL assay. At the time of enrolment, 18 patients fulfilled all the diagnostic criteria of PML [focal neurological disease with subacute progression; white matter lesions on MRI consistent with PML; detection of JCV DNA by PCR in the CSF (13 patients) or brain biopsy showing typical lesions of PML (five patients); no other likely aetiology]. These patients have proven PML (Cinque et al., 2003). Eight other patients fulfilled the clinical and neuroradiological criteria for PML, including focal neurological signs correlating with hyperintense lesions in T2-weighted images on MRI. Causes of leukoencephalopathy other than PML had been reasonably ruled out. However, PCR for JCV in the CSF was negative. No biopsy was performed in these subjects. We will refer to these patients as possible PML (Cinque et al., 2003). Among the 18 patients with proven PML, 13 were HIV+ and five were HIV. The eight patients with a possible form of PML were all HIV+.

Active versus inactive disease

To determine if the clinical evolution of PML was influenced by the presence of JCV-specific CTL, we tested the blood of patients as early as possible after the onset of PML and then followed them prospectively. Patients were divided into two groups based on the activity of PML as assessed by clinical and neuroradiological criteria. In some patients, the presence or absence of JCV DNA in the CSF was also determined during the course of the study. Those patients who developed an evolutive and aggressive disease were classified as having an ‘active’ form of PML, and those whose disease had stabilized or regressed were classified in the ‘inactive’ category (Cinque et al., 2003). Some of the PML patients who had a rapid fatal outcome after disease onset have been included in our retrospective studies (Koralnik et al., 2001; Du Pasquier and Koralnik, 2003; Du Pasquier et al., 2003).

Subset of patients with more than one CTL assay

To examine whether changes in the cellular immune response against JCV could herald alterations in the course of PML, peripheral blood mononuclear cells (PBMC) of 12 subjects were tested two to ten times (3.5 ± 2.5) for the presence of JCV-specific CTL, CD4+ T-cell count, HIV viral load and, in some of them, the presence of CTL specific for HLA-A*0201-restricted epitope on the Gag protein of HIV (HIV Gagp77-specific CTL) (see below).

Detection of JCV VP1-sp CTL

The methods used for detection of JCV VP1-sp CTL have been described in detail elsewhere (Koralnik et al., 2002; Du Pasquier et al., 2003). Briefly, two different assays were performed. In HLA-A*0201 subjects, JCV VP1-sp CTL were detected using a functional lysis assay. The effector T-cell response against VP1, the major capsid protein of JCV, was tested by stimulating PBMC of a given subject with pools of overlapping peptides encompassing the entire amino acid sequence of this protein. After 10–14 days of in vitro stimulation in the presence of recombinant human interleukin-2, these PBMC were tested in a 51Cr release assay using autologous B lymphoblastoid cell lines pulsed with the same pools of peptides as target cells (Du Pasquier et al., 2003). With this technique, any possible nonamer CTL epitope of JCV VP1 protein, could be detected. In HLA-A*0201+ subjects, JCV-specific CTL were detected in PBMC using HLA-A*0201/JCV VP1p36 or –p100 epitope tetramer staining assay. PBMC were tested for the presence of JCV VP1 peptide-specific CTL after 10–14 days of in vitro stimulation in the presence of peptide and recombinant human interleukin-2 as described previously (Koralnik et al., 2002; Du Pasquier et al., 2003). Of 26 subjects, 15 were HLA-A*0201+ (58%). These 15 patients had their PBMC tested with the tetramer staining assay, but PBMC of some of them were also tested by 51Cr release assay. We have shown in previous studies that the results of tetramer staining and 51Cr release assays are comparable for VP1p100 (r = 0.8) (Koralnik et al., 2002) and VP1p36 (r = 0.72) (Du Pasquier et al., 2003).

Detection of HIV Gagp77-specific CTL

To compare the CTL response against JCV and HIV, we tested the PBMC of HLA-A*0201+ subjects for the presence of CTL specific for an HLA-A*0201-restricted immunodominant epitope present on the Gag protein of HIV, the HIV Gagp77 epitope (Tsomides et al., 1994). This detection was performed by 51Cr release and by tetramer staining assays as described above.

Statistics

We used the median ± interquartile range (IQR) to report the ages, CD4+ T-cell counts/µl, HIV viral loads (copies/ml) and intervals between different time points in enrolled patients. The Fisher exact test and the Mann–Whitney test were used to compare the patients with versus those without JCV-specific CTL.

Results

To determine whether JCV-specific CTL are generated early in the course of PML and if their presence is predictive of subsequent disease activity, we performed a cross-sectional study evaluating the PBMC of PML patients as early as possible after disease onset. Of 26 patients, 15 had detectable JCV-specific CTL (58%). The interval between the first clinical manifestation of PML and the first CTL assay ranged from 1.0 to 9.6 months (3.9 ± 3.5, median ± IQR) in patients with detectable JCV-specific CTL and from 0.1 to 8.5 months (3.6 ± 1.9) in patients with undetectable CTL. This difference was not statistically significant (P = 0.80, Mann–Whitney). However, the period of follow-up between the first CTL assay and the end-point/death was significantly longer in patients with JCV-specific CTL (9.7 ± 5.0) than in those without JCV-specific CTL (1.8 ± 3.0, P = 0.02, Mann–Whitney). This is explained by the fact that most of the latter patients developed active disease and died a few months after the CTL assay was performed (Table 1).

View this table:
Table 1

Clinical and laboratory data on 26 study subjects

PatientHIVPMLHLA-A*0201HIV plasma viral load (copies/per ml)CD4+ T cells (/µl)JCV-specific CTLHIV-Gagp77-specific CTLOnset CTL 1CTL 1 end-pointPML course/outcome
1+pr.<50178+n/a7.79.6in/al
2+pr.1016104+n/a7.810.0in/al
3+pr.<50222+n/a9.37.9in/al
4+pr.+234320++4.021.6in/al
5+pr.+<50215+n/a1.711.7in/al
6pr.n/an/a+n/a3.914.7in/al
7pr.n/an/a+n/a4.42.4in/de
8+po.+60097++1.040.3in/al
9+po.+144204+1.514.2in/al
10+po.<50235+n/a9.512.7in/al
11+po.+198524++5.19.7in/al
12+po.+5530420++1.79.7in/al
13+po.+<50135++2.40.4in/de
14+pr.<5074+n/a1.90.4ac/de
15+po.+13 700273++2.35.6ac/al
16+pr.76631252n/a5.127.4in/al
17+po.+11 200196+8.57.8in/de
18+pr.51432n/a3.06.7ac/de
19+pr.n/a20n/a5.70.4ac/de
20+pr.+75 00045n/a4.50.1ac/de
21+pr.32596n/a4.33.0ac/de
22+pr.+n/a1n/a1.43.0ac/de
23+pr.+209 63417n/a0.11.3ac/de
24pr.+n/an/an/a1.31.8ac/de
25pr.+n/an/an/a3.31.6ac/de
26pr.+n/an/an/a3.60.6ac/de
  • Both 51Cr release and tetramer staining assays were performed on cells that had been stimulated with peptide and recombinant interleukin-2in vitro for 10–14 days. Onset CTL 1 is the time in months between the first neurological symptoms and the first CTL assay performed in a given patient; CTL 1 end-point is the time between the first CTL assay and the end-point represented by September 30, 2003 or death.pr. = proven; po. = possible; in = inactive; ac = active; al = alive; de = deceased; n/a = not available.

Patients with detectable JCV-specific CTL had an average age of 43.9 ± 6.0 years (median ± IQR) compared with 41.4 ± 6.4 years in patients without detectable JCV-specific CTL (P = 0.59). Among the HIV+/PML patients, the plasma HIV viral load was significantly lower in those with JCV-specific CTL (144 ± 431, median ± IQR) compared with those without (43 100 ± 54 778, P = 0.01). Similarly, the CD4+ T-cell count was higher in those who had detectable JCV-specific CTL (215 ± 103.5) than in those without this cellular immune response (32 ± 59.3, P = 0.004).

Patients with detectable JCV-specific CTL at the time of their first CTL assay

Among those patients who had detectable JCV VP1-sp CTL at the time of the first assay (patients 1–15, Table 1), 13 out of 15 (87%) developed an inactive form of PML. Among the 13 patients with inactive disease (patients 1–13), nine were still alive on September 30, 2003. Two additional HIV+ patients with proven PML were lost to follow-up 13.3 (patient 5) and 17.3 months (patient 1) after disease onset, respectively. At that time these patients were clinically stable. Finally, two patients who had an inactive form of PML died from unrelated causes. One, reported previously, was an 80-year-old HIV lady who had a history of Hodgkin's lymphoma three years prior to the onset of PML (patient 7). Almost 7 months after the onset of her neurological disease, she was improving neurologically but died from sepsis secondary to a urinary tract infection and decubitus ulcer (Du Pasquier et al., 2001). The other was an HIV+ patient with a possible form of PML who had a minimal and stable neurological deficit (patient 13). She died due to unclear circumstances during her sleep 2.7 months after disease onset. An autopsy was not performed. Finally, two patients had JCV-specific CTL and an active form of PML (patients 14 and 15). One patient presented with PML 3 weeks after the introduction of HAART, while her HIV viral load was decreasing and her CD4+ T-cell count was increasing (patient 14). Despite improvement of these immune parameters and despite the presence of a strong anti-JCV immune response mediated by CD8+ CTL, she had active PML and died 2.3 months after disease onset (Du Pasquier and Koralnik, 2003).

Patients with undetectable JCV-specific CTL at the time of their first CTL assay

Among the 11 patients who had no detectable JCV-specific CTL (patients 16–26), nine (82%) developed an active form of the disease (patients 18–26). All nine patients had proven PML and died 4.6 ± 1.4 months after disease onset. Two patients without detectable CTL had an inactive course of PML. One patient stabilized neurologically at the time the first CTL assay was performed, 5.1 months after disease onset (patient 16). This assay detected no JCV-specific CTL. However, JCV-specific CTL became detectable 3.6 months later (see below). The other patient was HIV+ with a possible form of PML (patient 17). A first CTL assay was negative 8.5 months after the beginning of neurological symptoms, but a second assay, performed 4.6 months after the first, was positive. He had an inactive form of PML but died 16.3 months after disease onset from hepatic failure due to HAART-related lactic acidosis. The fact that these two patients had no JCV-specific CTL at the first assay and yet had inactive disease might be due to other components of the immune response, such as JCV-specific CD4+ T cells. Alternatively, they might have been infected with a less virulent strain of JCV. Finally, although unlikely, a false-negative result on the first CTL assay of these two patients cannot be ruled out completely.

Thus, the early detection of JCV-specific CTL has an 87% (13/15) predictive value for an inactive form of PML, whereas the absence of such CTL has an 82% (9/11) predictive value for an active form of PML (P = 0.0009, Fisher exact t-test).

In addition, to ascertain the role of JCV-specific CTL in the earliest stage of PML, we analysed the results of a subset of 15 patients (4, 5, 6, 8, 9, 12, 13, 14, 15, 18, 22, 23, 24, 25, and 26; Table 1) who had a CTL assay performed less than 4 months after disease onset (1.9 ± 1.3 months). Among these patients, 7/9 (78%) with detectable CTL developed an inactive form of PML, whereas 6/6 (100%) of those without detectable CTL developed an active form of PML (P = 0.007). In this subgroup of patients, there was also no difference in the interval between the time of the diagnosis of PML and the first assay in those with detectable JCV-specific CTL (1.9 ± 0.5 months) and those without detectable JCV-specific CTL (2.2 ± 1.4, P = 0.69). These results further indicate that the emergence of JCV-specific CTL is seen very early in PML in those patients whose disease will become inactive.

Subset of patients with more than one CTL assay

To correlate the degree of PML activity and the profile of JCV-specific CTL over time, we performed multiple CTL assays on PBMC of a subset of patients. PBMC of 11 patients were tested two to 10 times for the presence of JCV-specific CTL (patients 1, 5, 8, 9, 10, 11, 12, 15, 16, 17 and 18; Table 2). These patients included eight who had JCV-specific CTL and three who did not have CTL detected at the time of their first assay. The lower number of patients in the latter category reflects their shorter survival (Table 1). The time interval between the first and the last assay in each individual ranged from 0.4 to 29.9 months (Table 2). Except for patients 5 and 8, who had six and 10 assays, respectively, the maximum number of assays was four. In addition to Table 2, patients 5 and 8 are described in case reports and in Fig. 2.

View this table:
Table 2

Patients with more than one CTL assay

PatientNumber of assaysMonths between first and last assayJCV-specific CTL (months elapsed since disease onset)aPML course/outcome
1234
122.3+(7. 7)+(10.0)n/an/ain/al
5*613.5+(1.7)+(4.0)+(6.0)+(9.9)in/al
8*1029.9+(1.0)+(3.4)+(7.2)+(10.3)in/al
934.5+(1.5)+(3.5)+(6.0)n/ain/al
10412.1+(9.5)+(12.9)−(14.5)− (21.6)in/al
1120.4+(5.1)+(5.4)n/an/ain/al
1221.5+(1.7)+(3.2)n/an/ain/al
1524.8+(2.3)+(7.1)n/an/aac/al
16316.7−(5.1)+(8.7)+(21.8)n/ain/al
1724.6−(8.5)+(13.1)n/an/ain/de
1825.9−(3.0)−(8.9)n/an/aac/de
  • a JCV-specific CTL were detected by tetramer staining with or without 51Cr release assays in HLA-A*0201+ subjects; PBMC of HLA-A*0201 subjects were tested by 51Cr release assay only.

  • * Detailed clinical, neuroradiological and immunological data on these two patients are given in the case reports (see text) and Figs 1 and 2. in = inactive; ac = active; al = alive, de = deceased.

Of the eight patients who had detectable JCV-specific CTL at the time of their first testing, four had positive results at all time points, and these results correlated with an inactive form of the disease (patients 1, 9, 11 and 12). A fifth patient (patient 5) was initially doing poorly clinically, but had a significant clinical improvement heralded by a pronounced increase in JCV-specific CTL. A sixth patient (patient 8) had JCV-specific CTL during the first 18 months of follow-up. When these cells disappeared, he presented with an episode of clinical and neuroradiological exacerbation, characterized by the occurrence of new PML lesions. The latter two patients are described in the case reports (below). A seventh patient (patient 15) had an active course of PML despite the presence of JCV-specific CTL. Her neurological symptoms occurred shortly after the introduction of HAART, probably representing PML in the context of an immune reconstitution inflammatory syndrome (Du Pasquier and Koralnik, 2003). Finally, an HIV+ patient (patient 10) with a severe but stable PML had two positive CTL assays initially, but 15 months after disease onset he lost the CTL response. Three months after CTL disappearance, the disease was noted to have progressed clinically.

Three patients had no JCV-specific CTL detected at the time of their first CTL assay. One HIV+/proven PML patient (patient 18) did not have detectable CTL in two out of two assays separated by a 6-month interval. His neurological disease remained active and he died 10 months after the onset of PML. Two patients had no detectable JCV-specific CTL at the time of their first assays, performed 5.1 and 8.5 months after PML onset respectively, but they had JCV-specific CTL on subsequent assays. The evolution of the disease was inactive in both. One of them (patient 16) remains alive almost 3 years after disease onset. However, the other one (patient 17), who had been diagnosed with possible PML, died 3 months after the second assay with hepatic failure due to HAART-related lactic acidosis. At autopsy, there were brain lesions consistent with PML, including bizarre astrocytes, lipid-laden macrophages, and positive staining in the nuclei of oligodendrocytes for polyomavirus by immunohistochemistry.

Patient 5 case report

This 45-year-old patient was diagnosed with HIV infection in 1995. He was not compliant with his HAART regimen. In December, 2000, he presented with a rapidly progressing static and kinetic cerebellar syndrome characterized by dizziness, gait imbalance, dysarthria, rotatory nystagmus and bilateral hand dysmetria. His plasma HIV viral load was 316 478 copies/ml and his CD4 count was 74/µl. A brain MRI showed hyperintense lesions in T2-weighted images in both middle cerebellar peduncles (Fig. 1, patient 5). The lesions did not enhance after gadolinium administration in T1-weighted images (not shown). The detection of JCV DNA by PCR was positive in the CSF, establishing the diagnosis of PML. HAART including lopinavir/ritonavir, d4T and ddI was started under close supervision in January 2001. In 1 month, his CD4+ T cells count rose to 176/µl (Fig. 2, patient 5, panel A) and his plasma HIV viral load dropped to 804 copies/ml (Fig. 2, patient 5, panel C). The patient nevertheless continued to worsen neurologically until late spring, 2001. However, in June, 2001, we observed a marked and relatively sudden improvement in his neurological condition, characterized by a decrease in hand dysmetria and the disappearance of nystagmus (Fig. 2, patient 5, vertical dotted line). His clinical condition remained stable until the end of January, 2002, when he left the country and was lost to follow-up.

Fig. 1

(Patient 5) Brain MRI performed in January 2001 shows hyperintensities in both middle cerebellar peduncles on T2-weighted images (arrows). (Patient 8) Brain MRI performed at the time of the clinical exacerbation in November 2001. There were several new hyperintensities, visible on T2-weighted images, (A) in the right pons and middle cerebellar peduncle, and (B) in the white matter of the left frontal lobe, as well as (C) an extension of a pre-existing lesion in the left thalamus (arrows).

Fig. 2

(Patient 5) An increase in JCV-specific CTL precedes clinical improvement of PML. After the introduction of HAART in January 2001, there was a rapid increase in CD4+ T-cell count (A) and a decline in HIV plasma viral load (C). JCV-specific CTL, which were almost undetectable by February 2001 (B), increased progressively during the following months (A and B). Interestingly, this increase started approximately 2 months before the clinical improvement. In contrast with the increase in JCV-specific CTL, there was a concomitant decrease in HIV Gagp77-specific CTL (A and B). (Patient 8) Disappearance of JCV-specific CTL heralds clinical exacerbation of PML. A decrease in JCV-specific CTL preceded the progression of PML (A′ and B′). Of note, the other immune parameters, such as the CD4+ T-cell count (A′) and HIV Gagp77-specific CTL (A′ and B′) remained elevated, and the HIV viral load (C′) became undetectable during this period. The vertical dotted line corresponds to a major clinical episode (improvement in patient 5, worsening in patient 8). The horizontal dashed line in panels A and A′ indicates the threshold (10%) above which the percentage of specific lysis is considered significant in the 51Cr release assay. The arrows at the bottom of the figure indicate when the brain MRIs displayed in Fig. 1 were performed for each patient. In patient 8, this arrow also indicates the time at which the PCR for JCV DNA was positive in the CSF, which confirmed the diagnosis of PML.

PBMC of this patient were tested six times over a 12-month period for the presence of JCV-specific CTL, starting in February, 2001, 1.7 months after the onset of his neurological symptoms. At this time, only a small percentage of his CD8αβ+ T cells stained with the HLA-A*0201/JCV VP1p100 tetramer, but the percentage of tetramer-positive cells rose over the following months, concomitantly with the first detection of JCV-specific CTL by the 51Cr release assay (Fig. 2, patient 5, panel B). In May of 2001, JCV VP1-sp CTL also became detectable using the 51Cr release assay. Interestingly, this rise in JCV-specific CTL preceded the clinical improvement which was seen in June, 2001 (Fig. 2, patient 5, panels A, B). Of note, while JCV-specific CTL responses were increasing, HIV Gagp77-specific CTL decreased (Fig. 2, patient 5, panels A, B), suggesting that the CTL responses against these two viruses were independent of each other.

Patient 8 case report

This 35-year-old-HIV-infected man stopped taking HAART in July 1999. In April 2000, he developed a left arm and face weakness and dysarthria. His HIV plasma viral load went up from 900 to 60 000 copies/ml, and his CD4+ T-cell count dropped from 150/µl to less than 100/µl. A brain MRI showed punctate hyperintensities in T2-weighted images in both thalami, in the left middle cerebellar peduncle, the right pons, and in both parietal white matters, predominating on the left. The clinical and radiological presentation was consistent with PML. However, a PCR for JCV in the CSF was negative. Other opportunistic infections or tumours were ruled out. He resumed HAART, including AZT, 3TC, indinavir and ritonavir and improved over the following months. By August 2000, a second brain MRI showed a decrease in the size of the thalamic lesions and disappearance of all the other lesions. In November, 2001 while he was compliant with HAART and had a CD4+ T-cell count greater than 150/µl, his neurological condition deteriorated dramatically. There was a new onset of cognitive dysfunction, diplopia and dysphagia, and worsening of dysarthria and dysmetria. Consistent with these findings, brain MRI showed an extension of pre-existing lesions and the appearance of new lesions in the white matter of the left frontal lobe, the right pons and the middle cerebellar peduncle (Fig. 1, patient 8, panels A–C). At that time, JCV DNA was detected by PCR in the CSF, which confirmed the diagnosis of PML (Fig. 2, patient 8, vertical dotted line). HIV RNA was undetectable in the CSF (<200 copies/per ml), but because his HIV plasma viral load was still detectable in the plasma (644 copies/ml), his HAART regimen was partially changed to AZT, d4T, efavirenz and lopinavir/ritonavir (Kaletra). After a few weeks the patient stabilized neurologically, but was left with severe neurological sequelae and a diminished quality of life compared with his condition prior to November, 2001. At the beginning of 2003, he moved to a different state and was followed by another neurologist. He remained stable until the endpoint of the study.

The PBMC of this patient were tested 10 times over a 30-month period for the presence of JCV-specific CTL. The first measurement was performed 1 month after the first neurological symptoms associated with his leukoencephalopathy. Repeated assessments of the cellular immune response showed disappearance of the JCV-specific CTL, as assessed by tetramer staining and 51Cr release assays (Fig. 2, patient 8, panels A′ and B′) shortly before the appearance of new lesions and the confirmation of his diagnosis of PML by a positive PCR for JCV DNA in the CSF. Contrasting with this decline in JCV-specific CTL, the CD4+ T-cell count and the number of HIV Gagp77-specific CTL remained high, indicating that his clinical worsening was not due to general impairment of his cellular immune responses. The plasma and the CSF HIV viral load remained low or were undetectable, indicating that HIV replication was well contained by HAART. One month after this progression of PML, in December, 2001, the JCV-specific CTL became detectable again, but only by tetramer staining assay (Fig. 2, patient 8, panel B′). One year later that they became detectable by 51Cr release assay (Fig. 2, patient 8, panel A′). Altogether, these results strongly suggest that the decline in JCV-specific CTL in November, 2001 allowed an increase in JCV replication and a flare of PML. This flare caused an exacerbation of his neurological condition, with the appearance of new PML lesions on MRI and a positive PCR for JCV in the CSF. After November, 2001, JCV-specific CTL became detectable again in his PBMC, correlating with a new stabilization of his neurological disease.

Discussion

The main goal of this prospective study was to evaluate the association between JCV-specific CTL in PBMC soon after the onset of PML and the subsequent containment of PML. In previous studies, we showed that the presence of JCV-specific CTL was associated with a favourable outcome in PML (Du Pasquier et al., 2001, 2003; Koralnik et al., 2001, 2002). However, in these studies, the majority of PML survivors had been diagnosed with PML more than 1 year prior to CTL testing, raising the possibility that the detection of JCV-specific CTL later in the course of the disease was only a reflection of an improvement of the patients' general immune status. The present results show that the detection of JCV-specific CTL within 4 months of the onset of neurological disease is predictive of a subsequent inactive course of PML in 85% of patients. By contrast, one can predict that the course of PML will be active in 82% of those patients without detectable CTL soon after the onset of clinical diseases (P = 0.0009). Moreover, when we analysed a subset of patients tested 2 months after disease onset, we found that the presence of JCV-specific CTL predicted inactive disease in 78% of patients, whereas the absence of such cells was indicative of an active form of PML in 100% of patients (P = 0.007). These results clearly show that the emergence of JCV-specific CTL is a very early event in PML and is a reliable prognostic marker of inactive disease.

In order to correlate the PML-related neurological events over time with the presence or absence of JCV-specific CTL, we performed repeated CTL assays in a subgroup of 11 patients. The correlation between the clinical evolution of PML and the profile of JCV-specific CTL is particularly well demonstrated in patients 5 and 8, who had six and 10 assays, respectively. Patient 5 had an increase in the number of JCV-specific CTL preceding a significant improvement of his neurological condition. This CTL increase was specific for JCV, as indicated by the fact that HIV Gagp77-specific CTL fell at the same time as JCV-specific CTL increased. This case suggests that the cellular immunity against JCV was not simply reflecting a global improvement of his general immune condition. Conversely, patient 8 had a dramatic decline in JCV-specific CTL that heralded a clinical exacerbation of PML. Interestingly, these events occurred while his general immune status, as reflected by his total CD4+ T-cell count, was stable. His anti-HIV immunity was preserved, as revealed by an HIV viral load that was low in the plasma and undetectable in the CSF. In addition, CTL specific for the immunodominant anti-HIV Gagp77 remained high throughout this period, demonstrating (i) that the JCV-specific immunity was independent from the HIV-specific immunity and (ii) that this JCV-specific CTL decrease was not due to impairment of CTL function in general.

HAART, by restoring general immune functions, is the only proven treatment for PML in patients who are HIV+. However, approximately 50% of HAART-treated patients still die from PML (Clifford et al., 1999; Gasnault et al., 2001). Our data demonstrate that the early appearance of JCV-specific CTL is the best available prognostic marker of a subsequent inactive course of disease. It is, however, likely that CD4+ T cells, which are present in higher numbers in HAART-treated patients, will provide more effective help for CTL than will CD4+ T cells from untreated HIV-infected patients (Rosenberg et al., 1997, 2000; Altfeld et al., 2001). As demonstrated recently, the help provided to CD8+ CTL by JCV-specific CD4+ T cells certainly plays a role in disease outcome (Gasnault et al., 2003). In this study, Gasnault and colleagues found a JCV-specific CD4+ T cell response in nine out of 10 PML patients with inactive disease after at least 6 months of disease evolution, whereas these cells were not detected in 14 PML patients with an active course of PML tested within 6 months of disease onset (Gasnault et al., 2001). However, since no data are available on the follow-up of these patients, the predictive value of JCV-specific CD4+ T cells early in the course of PML remains to be determined. Indeed, it is likely that some of these patients who had undetectable JCV-specific CD4+ T cells less than 6 months after PML onset will prove to have an inactive form of the disease, based on a previous study showing a favourable outcome in 53% of HAART-treated patients by this group (Gasnault et al., 2001).

Our data also shed new light on HIV+ patients with possible PML. Patients presenting with a disease clinically indistinguishable from PML, but with negative JCV PCR in the CSF, have become more frequent (Ammassari et al., 2000). In the present study, seven out of eight patients with possible PML had detectable JCV-specific CTL at the time of the first assay and six of these seven patients had an inactive course of PML. These results suggest that these patients may have a milder form of PML. In this context, the cases of patient 8 and 10 are particularly illustrative. These patients were diagnosed with possible PML at the time of enrolment. PML was proven later in patient 8, when the detection of JCV DNA by PCR in the CSF became positive shortly after the decline in JCV-specific CTL. Patient 10 had an inactive form of possible PML and died from an unrelated condition. At autopsy, the brain exhibited typical lesions of PML, yet the detection of JCV DNA by PCR had been negative in the CSF studies. Based on our data, we can infer that the prevalence of PML is certainly underestimated. Since seven out of eight patients with possible PML had an inactive course of the disease, our results also indirectly confirm the fact that the prognosis of PML is inversely correlated with the level of JCV DNA in the CSF as assessed by PCR (Koralnik et al., 1999).

The temporal association between containment of PML and the emergence of JCV-specific CD8+ CTL suggests that JCV-specific CTL are crucial in the early control of PML. CD8+ T cells have been shown to be of central importance in containing the replication of a variety of viruses. In HIV-infected humans, the emergence of HIV-specific CD8+ T cells occurs early after viral infection (Borrow et al., 1994; Koup et al., 1994; Pantaleo et al., 1994; Wilson et al., 2000) and coincides with the decrease in HIV viraemia (Ogg et al., 1998; Wilson et al., 2000). Studies in the simian immunodeficiency virus (SIV)-infected monkey animal model confirmed that CD8+ CTL played a crucial role in the containment of this virus (Jin et al., 1999; Kuroda et al., 1999; Schmitz et al., 1999). Virus-specific CD8+ CTL also play an important role in controlling herpesvirus replication. In the case of the γ-herpesvirus Epstein–Barr, which can cause both a latent and a lytic infection in humans, Epstein–Barr virus-specific CD8+ CTL appear as early as in the first week after the initial manifestation of infectious mononucleosis (Steven et al., 1996; Catalina et al., 2001) and are important in controlling the viraemia (Callan et al., 1998). Cytomegalovirus is a human β-herpesvirus which establishes a persistent infection in humans and, like JCV, can cause severe morbidity in immunosuppressed individuals (Mocarsky and Courcelle, 2001). Clinical cytomegalovirus-related disease occurs in those bone marrow transplant recipients who have a delay in the regeneration of the CD8+ T cell response (Li et al., 1994). Our data indicate a firm association between the early presence of JCV-specific CD8+ CTL in peripheral blood and the containment of PML. Studies addressing specifically the role of these cells in the CNS of PML patients are warranted.

Acknowledgments

This work was supported by a Public Health Service grant RO1 NS/AI 41198 and NS 047629, a grant from The Harvard Center for Neurodegeneration and Repair, and the Dana Farber Cancer Institute-Beth Israel Deaconess Medical Center-Children's Hospital Center For AIDS Research grant P30-AI28691 to I.J.K. R.A.D.P. is the recipient of a fellowship for advanced researchers from the Swiss National Science Foundation and a grant from the Eugenio Litta Foundation. We are grateful to Joern Schmitz, Michelle Lifton and Darci Gorgone, who run the flow cytometry core facility of our own.

Footnotes

  • * Present address: Division of Neurology and Division of Immunology, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland

References

View Abstract