This scientific commentary refers to ‘JC virus spread is potentiated by glial replication and demyelination-linked glial proliferation’ by Li et al. (https://doi.org/10.1093/brain/awae252).
The JC virus (JCV) is a ubiquitous human pathogen that typically results in lifelong, asymptomatic infection. In individuals with profound immune compromise, historically often due to HIV infection, JCV can cause a devastating demyelinating disease of the CNS known as progressive multifocal leukoencephalopathy (PML). PML has a high fatality rate and poses a significant challenge due to the absence of validated therapies. Although effective antiretroviral therapy has decreased the occurrence of HIV-associated PML, medical advances in oncology, transplant medicine and immunology have resulted in longer survival of immunodeficient patient populations. Consequently, PML has re-emerged as a much-feared complication of biological therapies targeting immune pathways.
In recent years, there has been a paradigm shift in the treatment of PML with increased focus on re-establishing an adaptive antiviral response against JCV. This has led to a surge of reports describing immunotherapeutic strategies for PML, including checkpoint inhibitors,1 recombinant interleukin-7,2 and virus-specific T cells.3 However, success has been inconsistent, likely due to factors such as variability in specific cellular target numbers, severely compromised bone marrow or thymic function, genetic immune deficiencies, and the risk of treatment-limiting adverse immune reactions driven by underlying conditions. As such, effective immunotherapeutic interventions will likely require detailed knowledge of an individual patient’s immunological status to best tailor personalized treatment approaches.
Improved understanding of host–viral interactions, and of mechanisms underlying JCV-associated cellular injury, may uncover new therapeutic targets with broader applicability and improved outcomes. However, a significant barrier to progress in this area has been the highly restricted host range and selective cell tropism of JCV, which have impeded the development of suitable model systems. The lack of an animal model, limited availability of primary human brain cells, and difficulty in establishing an efficient in vitro cell system have hampered research in JCV.
The JC virus was first isolated in 1971 following inoculation of primary human fetal glial cells with extracts from the brain of a patient who had succumbed to PML. Initial studies of JCV required use of primary glial cell cultures, but development of cell lines transformed with SV40 or JCV large T antigen (LTAg) subsequently allowed more efficient propagation of JCV in vitro. Among the most widely used culture models is the SVG-A cell line, a human fetal astroglial cell line transfected with an origin-defective SV40 LTAg. Another approach uses infection with a chimeric polyomavirus, JCV (Mad 1/SVEΔ), in which the non-coding region of JCV is replaced with that of SV40 to enhance replicative efficiency.4
While these systems overcome challenges of culturing JCV in vitro, they do not reproduce the physiological conditions of human infection. In fact, it is notable that none of these culture systems include oligodendrocytes, the primary cellular target in the brains of patients with PML. Oligodendrocytes are difficult to sustain and grow in culture and there are few established cell lines that support JCV replication. By contrast, astrocytes and glial precursors represent a more stable and reproducible culture environment, and have been instrumental in studying the viral life cycle and testing candidate antiviral compounds.
Over the years, important progress has been made in understanding JCV pathophysiology, including the discovery that intra-host viral evolution within the non-coding control region of JCV leads to the emergence of pathogenic variants by altering transcription factor binding sites. It is these rearrangements, rather than differences in viral binding or entry, that help explain why only specific cell types support JCV replication.5
The unique molecular virology of JCV has impeded attempts to establish an animal model. JCV does not replicate in non-human hosts, and inoculation of mice or hamsters results in tumours rather than reproducing PML pathology such as demyelination. Likewise, while studies using murine polyomavirus have provided insights into host immune responses, the pathology of tumours does not model PML. Transgenic mouse models expressing LTAg in oligodendrocytes can mimic some dysmyelinating phenotypes due to suppressed myelin basic protein expression, but do not capture the full spectrum of PML.6 Immune-deficient mouse models with human lymphocyte engraftment7 do not develop CNS disease, likely due to the absence of human brain tissue. Rhesus monkeys infected with SV40 develop PML-like disease in only about 2% of cases, and thereby do not provide a reliable model for study.6
Kondo and colleagues8 took a novel approach by engrafting human glial precursor cells (hGPC) into congenitally hypomyelinated (Rag2−/−/Mbpshi/shi) mice to create a human glial chimeric mouse model. Intracerebral injection of these mice with JCV led to infection of astrocytes and hGPC, with oligodendrocyte loss and demyelination occurring later. The death of infected oligodendrocytes was attributed to apoptosis triggered by forced cell cycle entry due to LTAg, rather than direct cytolytic infection. Using a second human glial chimeric mouse that lacked human oligodendrocytes, the group then showed that oligodendrocytes were not required for JCV propagation. Based on these findings, they concluded that oligodendrocyte loss and demyelination are secondary events in PML.
Building on this previous work, in this issue of Brain, Li and co-workers9 explore JCV–host interactions using primary fetal astrocyte cultures and another glial chimeric model, this time based on Rag1−/− mice engrafted with induced pluripotent stem cell (iPSC)-derived glial precursor cells. In cell cultures, they show that JCV selectively replicates in proliferating glia, and that JCV infection itself induces glial proliferation. Remarkably, they also observed induction of proliferation in uninfected bystander cells, both as a reaction to demyelination but also in response to signalling from infected astrocytes. Transcriptomic analysis in these mouse models revealed a complex feedforward mechanism of JCV propagation involving paracrine signalling. A DNA damage response in infected cells induces inflammatory signalling, leading to bystander proliferation and injury, which in turn amplifies and accelerates the spread of the virus. As in previous models, astrocytes and glial precursor cells were the primary infected cell types, while a large proportion of dying oligodendrocytes were uninfected bystanders (Fig. 1).
Cellular injury in PML. (A) Feed-forward model of cellular injury in PML proposed by Li et al.9 (B–D) Histopathologic features of PML. (B) Foci of demyelination (Luxol fast blue) corresponding to (C), regions of abundant productive JCV infection as detected by VP1+ immunohistochemistry, with prominent staining of oligodendrocyte nuclei (inset). (D) Typical oligodendrocyte with ground glass nucleus (haematoxylin and eosin), reflecting productive JCV infection. Image courtesy of Jaskaran Grewal.
This work provides valuable insights into host–virus interaction during JCV infection and the mechanisms of cellular injury, suggesting potential new treatment approaches targeting innate immune response pathways.
When translating these findings to human disease, it is critical to note that unlike in human glial chimeric models, mature oligodendrocytes support JCV replication in PML, as revealed by the presence of abundant viral particles in these cells, whereas this is less prominent in astrocytes. The route of JCV entry—via direct inoculation into the corpus callosum—as well as the nature of the immune deficiency in this model have been suggested to contribute to the observed phenotypic differences.
Recently, another group10 reported successful JCV infection of a human 3D brain model generated from iPSCs. In this model, the majority of infected cells were O1+ and NOGOA+ oligodendrocytes, which contained JCV VP1 and LTAg-positive inclusions. JCV VP1 inclusions were present in astrocytes to a lesser extent. This fully human culture system thus appears to more closely recapitulate classic features of PML. Future work will need to bridge the gaps between these model systems and further clarify the mechanisms of JCV pathogenesis in PML, to accelerate development of much needed treatments.
Funding
I.C. is supported by the NINDS Intramural Research Program; C.S.T has received funding from NIH R01NS116278.
Competing interests
I.C. is a shareholder of Keires, AG; Nouscom, AG; and PDC*line pharma, outside this submitted work. C.S.T. reports no competing interests.
