Progressive multifocal leukoencephalopathy (PML) is a fatal demyelinating disease of the central nervous system (CNS) caused upon reactivation of the human neurotropic polyomavirus, JC virus (JCV), in patients with impaired immune systems. Often, PML is associated with lymphoproliferative disorders, antineoplastic therapy for myeloproliferative disorders and cancer, sarcoidosis, tuberculosis, Whipple's disease, autoimmune diseases, and non-tropical sprue (Brooks and Walker, 1984). PML, a previously rare complication of middle-aged and elderly patients with lymphoproliferative disorders, is now commonly seen in patients of different age groups due to the increasingly widespread use of immunosuppressive chemotherapy and the prevalence of the acquired immune deficiency syndrome (AIDS) (Walker, 1985; Berger et al., 1987).
Examination of PML brains shows plaque-like lesions that are most frequently found in the deep cortical layers at the gray–white junction and within the white matter itself (Richardson, 1961). Early in the disease, multiple small plaques of demyelination, ranging up to several millimeters in diameter, can be found in the white matter. The clinical progression of PML appears to result from the enlargement and confluence of these early plaques into larger lesions, which may become necrotic. Histopathology suggests that PML is a disease that primarily targets oligodendrocytes and to a lesser extent astrocytes (ZuRhein, 1969). The white matter lesions result from the destruction of the oligodendrocytes with associated myelin degeneration. A second histologic feature characteristic of PML is the presence of giant multinucleated astrocytes. These bizarre cells may contain multiple heterochromatic nuclei and display abnormal mitotic figures (see Fig. 1). These cells are morphologically indistinguishable from the giant astrocytes present in pleomorphic glioblastomas (Astrom et al., 1958; Richardson, 1961). As the pathological features of PML became more widely recognized, electron microscopic evidence demonstrated polyomavirus particles in the enlarged nuclei of oligodendrocytes containing inclusion bodies (ZuRhein and Chou, 1965; ZuRhein, 1967, 1969). It is now well established that a ubiquitous human polyomavirus, JCV, is the causative agent of PML (ZuRhein, 1967; Johnson, 1982; Walker, 1985).
Infection with JCV appears to be a common event during childhood, yet nothing is known about the primary infection. In immunocompetent hosts, JCV elicits neutralizing antibodies which inhibit the natural ability of JCV to agglutinate human type O erythrocytes (Gardner, 1973; Mantyjdurvi et al., 1973; Padgett and Walker, 1973). Sixty-five to ninety-two percent of the population express serum antibodies by 17 years of age, however, have never exhibited symptoms of clinical disease that could be ascribed to JCV. In general, it is accepted that JCV, like many human viruses is capable of establishing latency and that PML arises from reactivation of JCV from a latent state (Chesters et al., 1983; Major et al., 1992a). Probable sites of latency are kidney cells and B lymphocytes (Houff et al., 1988; Major et al., 1992a). JCV is detected in the urine of renal transplant patients and pregnant women (Coleman et al., 1980, 1983; Hogan et al., 1980; Gardner et al., 1984) and JCV DNA can be detected by PCR in the kidney of non-immunocompromised individuals (Loeber and Dörries, 1988; Yogo et al., 1990; Flaegstad et al., 1991), arguing for kidney cells as a site of JCV latency. JCV DNA can also be detected in B lymphocytes in 75–85% of bone marrow, spleen, and brain from PML patients (Ho et al., 1984; Major et al., 1992b; Tornatore et al., 1992). Circulating peripheral blood lymphocytes infected with JCV are found in greater than 95% of PML patients' blood (Dr. Sidney Houff, personal communication). Activated B cells can cross the blood-brain barrier and may transmit JCV to oligodendrocytes within the brain (Houff et al., 1988). Whereas kidney and B cells may provide a site for JCV latency, JCV replicates efficiently only in cells of glial origin (Padgett and Walker, 1976; Dorries, 1984). However, it is still a matter of controversy as to whether JCV is latent in glial cells (Lipton, 1991). Some studies have failed to detect JCV DNA in brains of healthy individuals and argue that JCV remains latent outside the CNS (Telenti et al., 1990). There is one report of JCV DNA within the brains of elderly individuals (Mori et al., 1991) and one report of low levels of JCV DNA within the brains of non-immunocompromised individuals (White et al., 1992) which implicates the brain as a potential site for JCV latency. In any case, JCV appears to be present in a high percentage of normal adults at levels sufficient to maintain specific neutralizing antibody titers in the circulation.
The appearance of PML is associated with the breakdown of immunocompetence indicating that immune system function may have an impact in reactivation of viral gene expression and replication of JCV. Since patients with PML still have detectable anti-JCV serum antibodies (Johnson, 1982; Walker, 1985) and detectable anti-JCV antibodies in the cerebral spinal fluid (CSF) (Guillaume et al., 2000), the immune impairment relevant to this disease is most likely in the cell-mediated defenses. Most PML patients exhibit anergy to skin tests with common recall antigens (Marriott et al., 1975; Mathews et al., 1976; Varakis et al., 1978; Dorries, 1984) and impaired T cell proliferative responses to alloantigens and to mitogenic stimulation with PHA (Knight et al., 1972; Marriott et al., 1975; Wobbe et al., 1985). Additionally, while mitogen induced production of lymphokine migration inhibitory factor (LIF) is normal in PML patients tested, the production of LIF in response to JCV antigens is absent (Willoughby et al., 1980). While MHC class 1 and II molecules are expressed at high levels within PML lesions (Achim and Wiley, 1992), their presence alone does not appear to stimulate an appropriate immune response to combat JCV in most cases of PML. Recent studies of long-term survivors of PML indicate the cytolytic activity mediated by class I restricted CD8 + T cells against either T antigen or the VP1 capsid protein is associated with either recovery from or lack of progression of PML (Du Pasquier et al., 2001; Koralnik et al., 2001). The peripheral blood mononuclear cells (PBMCs) from PML survivors lysed B-lymphoblastoid cells lines expressing either T antigen or VP1 protein, while PML progressors PMBCs were not cytolytic. Further, a nine amino acid epitope of VP1 (ILMWEAVTL) was discovered which was recognized by the CTL of PML survivors who were also HIV positive and HLA-A2 positive (Koralnik et al., 2002).
A recent study, which examined both the humoral and cellular immune response due to PML disease, found that the proliferative capabilities of PBMCs from PML patients and from PML/AIDS patients was suppressed compared to healthy controls (Weber et al., 2001). Additionally, IFNγ production from VP1 stimulated PBMCs of PML and PML/AIDS patients was suppressed compared to normal controls. IFNγ causes a switch from the TH1 subset of Th cells, which is responsible for cell-mediated immunity to the TH2 subset of Th cells, which is necessary for B cell activation and antibody production. These data agree with the earlier data showing high antibody levels in persons infected with JCV and a low cell immunity response in PML patients. Active JCV infection can cause a perturbation in cytokine levels, leading to a strong antibody response while dramatically reducing the cellular immunity necessary for destroying the virus. It was also shown that PBMCs from patients with PML and AIDS had an increase in IL-10 production in response to VP1 protein compared to normal controls, PML patients or AIDS patients, which further drives the switch to production of TH2 cells. This indicates that the dual activities of HIV and JCV cause more fluctuations in the human immune system than either virus alone.
Similar to other polyomaviruses, JCV possesses a circular genome of double-stranded DNA within an icosahedral capsid. The prototype strain of JCV, Mad-1, contains 5130 nucleotides (Frisque et al., 1984), which can be functionally divided into three regions; an early coding region, a late coding region, and a regulatory (non-coding) region (for review see Raj and Khalili, 1995) (Fig. 2). The regulatory region, which contains the promoter/enhancer for early and late gene transcription as well as the origin of DNA replication is located between the early and late coding regions. The viral early genes encode the viral regulatory protein, T-antigen, whereas the late genes encode the structural capsid proteins VP1, VP2, and VP3 as well as Agnoprotein (Raj and Khalili, 1995). The viral lytic cycle begins by expression of the viral early protein, T-antigen, which occurs before replication of viral DNA. T-antigen is a multifunctional protein which interacts with several host regulatory proteins and, by manipulating host gene expression and/or their function, orchestrates subsequent steps of the viral life cycle including viral DNA replication and activation of late gene transcription. The products of the late genes, the capsid proteins, accumulate in the nucleus and associate with the replicated viral DNA forming virions which, in turn, lyse the host cells. Thus, T-antigen acts as the central regulator of the viral lytic cycle. To exert its regulatory action on both DNA replication and gene transcription, JCV T-antigen requires the participation of host factors.
As JCV displays a narrow species and host cell tropism, which restricts its productive replication to human brain cells, many laboratories, including our own, have focused their attention on deciphering the molecular basis for host cell restriction of JCV and identifying the participant factors (for review see Frisque and White, 1992; Raj and Khalili, 1995). Comparison of JCV expression in CNS cells (glial origin cells) and non-glial cell types has demonstrated that JCV early promoter activity is significantly higher in glial cells (Kenney et al., 1984; Feigenbaum et al., 1987; Lashgari et al., 1989; Tada et al., 1989) suggesting that the promoter/enhancer region of JCV acts in a cell type-specific manner. Because activation of the early genes is essential for productive infection, non-glial tissues which cannot support expression of the JCV early genome (i.e., large T-antigen) fail to stimulate viral DNA replication or late gene transactivation. In support of this hypothesis, earlier studies have indicated that JCV DNA replication and late gene transcription occur in non-glial cells which contain an exogenous source of the JCV early gene product (Feigenbaum et al., 1987; Lashgari et al., 1989). These experiments suggest that the host cell restriction of JCV in humans is predicated on the ability of glial cells, but not non-glial cells, to produce transcription factors which transactivate early gene expression resulting in T-antigen expression.
The cell type restriction of JCV early gene expression has been confirmed in vivo. Transgenic mice, containing JCV T-antigen under the control of the JCV early promoter, express T-antigen in cells derived from neural origin (Small et al., 1986). Additionally, transgenic mice containing chimeric viral early sequence in which the control regions from SV40 and JCV were exchanged also demonstrated the specificity of the respective viral regulatory regions (Feigenbaum et al., 1992). All of these observations indicate that the neurotropism of JCV is based upon the glial-restricted expression of the viral early genes.
The observation that the non-coding regulatory region of JCV is responsible for the host range and glial cell tropism attracted the attention of many laboratories to perform genetic and molecular biology studies to identify the important cis-acting regulatory elements within the viral promoter and define the trans-acting cellular factors from CNS cells that recognize specific sequences within the regulatory region of JCV and induce viral early and late gene expression. The regulatory region of JCV contains the origin of viral DNA replication and several transcriptional control elements. These regulatory elements provide a target for several ubiquitous and cell type-specific DNA-binding transcription factors. Interestingly, several of these regulatory factors can be induced by signaling pathways upon treatment of cells with immunomodulators and cytokines. For example, the JC virus promoter contains several regulatory motifs which are responsive to various inducible transcription factors including nuclear factor κB (NFκB) (Ranganathan and Khalili, 1993; Raj and Khalili, 1995; Mayreddy et al., 1996; Safak et al., 1999). Interestingly, tumor necrosis factor alpha (TNFα), a well known inducer of NFκB activity, was detected in brain specimens of a patient with PML, suggesting a potential role for NFκB-induced gene transcription in the pathogenesis of PML (Nagano et al., 1994). This element may thus modulate the activity of JCV during bouts of immunosuppression when cytokine levels are altered. In support of this role, previous studies have shown that the JCV NFκB sequence is responsive to TNFα and interleukin-2 (Mayreddy et al., 1996).
The significantly higher incidence of PML in AIDS patients than in other immunosuppressive disorders has suggested that the presence of HIV-1 in the brain may participate, directly or indirectly, in the pathogenesis of this disease. In support of this concept, previous studies have demonstrated that the HIV-1 encoded trans-regulatory protein, Tat, greatly increases the rate of transcription from the JCV late promoter (Chowdhury et al., 1990, 1992, 1993). Tat is a potent HIV-1 derived regulatory protein produced at an early phase of infection and plays a key role in HIV-1 replication. This protein is expressed during the early phase of the HIV-1 life cycle where it serves to activate transcription directed by the HIV-1 LTR (Cullen, 1990; Frankel, 1992). Tat activation leads to increased expression of transcripts encoding all viral proteins, including Tat itself, resulting in a positive feedback cycle and a massive induction of viral gene expression. While the mechanism of Tat activation has yet to be fully elucidated, a great deal has been learned thus far. Unlike classical transcription factors which bind DNA, Tat binds to a specific RNA structure, termed the TAR element, found at the 5′ end of all HIV-1 mRNAs (Fig. 3). Experiments carried out in a limited number of cell types, primarily HeLa cells and T lymphocytic cell lines, indicate that the interaction of Tat with this specific RNA target is an important component of the activation mechanism. Accordingly, the TAR element has been referred to by some as an “RNA enhancer” (Berkhout et al., 1989, 1990; Jeang and Berkhout, 1992). It has been suggested that binding to TAR serves to localize Tat to the vicinity of the promoter where it interacts with unspecified targets resulting in enhanced transcription. A number of reports have suggested that the primary function of Tat is to increase the efficiency of transcriptional elongation (for review see Laspia et al., 1989, 1990; Feinberg et al., 1991; Karn, 1999). The ability of Tat to influence elongation seems surprising in light of evidence that the target of Tat function is at the promoter, and the kinetics of Tat function indicate an effect at the time of, or immediately after, initiation of transcription. In an attempt to resolve this apparent paradox, it has been proposed that Tat functions at the promoter of HIV-1, requiring the participation of upstream regulatory elements, and biases initiation towards the formation of “more-processive” transcription complexes.
The mechanism by which Tat may directly exert its transcriptional activity on the heterologous JCVL (JCV late) promoter was investigated by identifying the cis-acting Tat-responsive element within the JCVL promoter. The control region for the JCVL promoter is comprised of two 98 bp tandem repeats which confer glial-specificity to viral gene transcription (Tada et al., 1990). There are multiple late transcription units that are transcribed from various regions spanning the 98 bp repeat sequence (Major et al., 1992a). Computer-assisted analysis of the JCV control region has revealed substantial homology between the control region of the 98 bp repeat and the HIV-1 TAR sequence. This homology resides in the leader of some JCV late RNA species. Results from site-directed mutagenesis, as well as TAR-swapping experiments revealed that the TAR homology of the JCVL promoter is responsive to HIV-1 Tat induction and may participate in the overall activation of the JCVL promoter mediated by this transactivation (Chowdhury et al., 1992) (Fig. 3). The preferential enhancement of JCVL promoter activity by Tat in glial cells suggests that in addition to the TAR homologous sequence, a secondary target that includes a tissue-specific element(s) may participate in glial-specific induction of the JCVL promoter by Tat. In this respect, a secondary target sequence enriched in GA/GC located upstream of the JCVL RNA initiation sites termed upstream tar (upTAR) that positively responds to Tat activation in glial cells was identified (Chowdhury et al., 1993) (Fig. 3). The cellular protein, Purα, binds to this upTAR region along with HIV-1 Tat and together increase transcription directed by the JCVL promoter and viral replication (Krachmarov et al., 1996; Wortman et al., 2000; Daniel et al., 2001). Perhaps it should be noted that HIV-1 infected cells are capable of secreting viral regulatory proteins, including the HIV-1 transactivation protein, Tat (Frankel and Pabo, 1988; Ensoli et al., 1990). Once released, it can diffuse into neighboring uninfected and/or infected cells and alter expression of the responsive cellular and resident viral genes. Thus, reactivation of the JCVL promoter by HIV-1 may not require co-infection of glial cells with these two viruses.
In addition to its direct effect, HIV-1 Tat may indirectly, most likely through cytokines, participate in de-regulation of cellular and activation of the heterologous viral promoter in neural cells. There is a large body of evidence demonstrating that the CNS utilizes cytokine cascades in normal physiologic and pathologic states (for review see Merrill, 1992). These cytokines can originate from resident cells of the nervous system, specifically astrocytes and microglial cells or from cells of the immune system that traffic through the CNS (Persidsky et al., 1999). The targets of the cytokines within the CNS may include cells of both neuronal and glial origin (Gallo et al., 1989). Several cytokines, such as TNF, interleukin 1 (IL-1), and transforming growth factor β (TGFβ) may exert toxic effects on CNS cells, and have been postulated to contribute to the pathogenesis of the neurological complications of HIV-1 infection (Grimaldi et al., 1991; Merrill and Chen, 1991; Genis et al., 1992; Merrill, 1992; Tyor et al., 1992, 1993). In this respect, one can envision a scenario in which activated/infected macrophages and other microglia, by secreting cytokines, may attract other inflammatory cells to the CNS, and stimulate astrocytes and endothelial cells to produce other soluble factors. These soluble factors (including cytokines), in turn, contribute to CNS pathology, either by acting on uninfected neighboring cells or by stimulating the production of active viruses (HIV-1, JCV, etc.) from persistently infected glial cells (Fig. 4). In support of this model, it has been demonstrated that during AIDS, expression of two important modulators, IL-1 and TNFα, is increased in the brain (for review see Tyor et al., 1992). These two cytokines have a strong capacity to increase transcription of HIV-1 and other cellular promoters through the κB enhancer element (Israel et al., 1989; Osborn et al., 1989; Tornatore et al., 1991). Of particular interest, earlier studies by Major et al. (1990) revealed the presence of a putative κB sequence on the early side of JC viral origin of DNA replication. κB is a potent transcription enhancer which was first identified in the kappa light chain of the immunoglobulin gene, and later in many other cellular and viral transcription units (for review see Bauerle, 1991; Ghosh et al., 1998; Pahl, 1999). The transcription module has been shown to confer inducible transcriptional activity when placed upstream of a reporter gene, and two tandem copies of the element greatly bolster this effect. This enhancer binds to a growing family of proteins called nuclear factor κB (NFκB). NFκB is a complex of two proteins, the most common of which are p50 and p65, that is believed to bind as a heterodimer. These proteins exist in the cytoplasm of most cells in an inactive form complexed to an inhibitor termed IκB. Stimulation by a number of agents such as PMA, lipopolysaccharide (LPS), or TNFα, results in the dissociation of the IκB-NFκB complex, by protein kinase C-dependent phosphorylation of IκB (for review see Bauerle, 1991; Karin, 1999). The unbound NFκB is, in turn, transported to the nucleus where it binds to its target sequence and exerts its activity. Previous studies in our laboratory have suggested that the 10 nucleotide κB sequence homologous to the classical κB element found in immunoglobulin and HIV-1 positively influences expression of the JCV promoter in human glial cells (Ranganathan and Khalili, 1993). Results from binding studies have revealed the association of multiple inducible and non-inducible nuclear proteins from glial cells that differentially bind to JCV κB and the classical κB sequence (Fig. 4). However, the importance of this regulatory motif and its interactive proteins in the physiology of the JCV life cycle in normal cells and cells treated with a variety of cytokines and other regulators such as TNFα, IL-1, LPS, etc., remains to be investigated.
In addition to TNFα and IL-1, it has been demonstrated that the level of TGFβ, a potent suppressor of the immune system is significantly increased within the brain in response to infection with HIV-1 (Wahl et al., 1991). In vitro cell culture studies have indicated that HIV-1 Tat is capable of increasing expression of TGFβ in human astrocytic glial cells (Houff et al., 1988; Cullen, 1990; Lotz et al., 1990). Given that Tat up-regulates expression of TGFβ, two potential consequences deserve consideration: First, the induced TGFβ may act in a paracrine fashion, altering the expression of the extracellular matrix (Rossi et al., 1988). Accordingly, earlier observations suggested an increase in expression of the extracellular matrix protein, collagen, α1 type I and α1 type III, as well as fibronectin, in stable cell lines expressing the Tat protein (Taylor et al., 1991). Thus, it is likely that Tat, by elevating the level of TGFβ in the cells, indirectly modulates expression of the matrix genes. With the notion that TGFβ acts as a chemotactic agent for monocytes (Wahl et al., 1987), it was postulated that TGFβ present in the brain of HIV-1 infected individuals may facilitate recruitment of other infected monocytes into the brain, thereby perpetuating CNS dysfunction (Tada et al., 1990). TGFβ may also function in an autocrine fashion, as increasing HIV-1 expression would be expected to subsequently activate TGFβ through Tat. Several studies have demonstrated the ability of TGFβ to upregulate HIV-1 replication in macrophages. In general, the TGFβ family members initiate signaling from the cell surface by binding to a heteromeric complex of two distinct but related serine/threonine kinase receptors. Binding of the ligand to the type II receptor results in the recruitment and phosphorylation of the type I receptor. This activates the type I receptor, which propagates the signal to a family of intracellular signaling mediators known as SMADs (Heldin et al., 1997).
SMADs are a novel family of signal transducers that have been implicated as downstream effectors of TGFβ signaling. These proteins move into the nucleus, target specific genes and generate transcriptional complexes of specific DNA-binding ability (Zhang et al., 1996; Massague, 1998). In the basal state, SMADs exist as homo-oligomers that reside in the cytoplasm. Upon ligand activation of the receptor complex, the type I receptor kinase phosphorylates SMAD2 and SMAD3, which then form a complex with SMAD 4 and translocates into the nucleus, where these complexes, either alone or in association with a DNA-binding subunit, activate target genes by binding to specific promoter elements (Fig. 4).
In addition to these well-studied pathways, inflammatory cytokines and immunomodulators may exert their effect upon viral gene expression by activating or suppressing other groups of DNA binding proteins such as GBPi. GBPi is a distinct inducible DNA binding protein that binds specifically to the double-stranded GGA/C-rich region of the JCV control region within the origin of viral DNA replication (Raj and Khalili, 1994). The induction of GBPi is widespread and mediated by many inflammatory cytokines including IL-1β, TNFα, IFNγ, and TGFβ. Interestingly, the induced protein acts as a transcriptional repressor in its native context in the JCV promoter, although when its binding sequence is transposed to a heterologous promoter, GBPi appears to function as a transcriptional activator. These observations suggest a role for GBPi in cytokine-mediated induction of the JCV genome and support a model for its dual function as an activator and suppressor which depends upon its interaction with other DNA binding proteins associated with its neighboring regulatory motifs.