The Demyelinating Dilemma: Progressive Multifocal Leukoencephalopathy
Progressive multifocal leukoencephalopathy (PML) has long remained one of the most daunting challenges in clinical neurology and infectious disease. Caused by the reactivation of the JC polyomavirus (JCV) in the setting of severe cellular immunodeficiency, this opportunistic infection leads to the lytic destruction of oligodendrocytes, resulting in rapidly progressive demyelination of the central nervous system. Historically, PML carried a dismal prognosis, particularly in patients with HIV/AIDS, hematologic malignancies, or those receiving potent immunosuppressive therapies like natalizumab for multiple sclerosis.
Until recently, the therapeutic cornerstone for PML was the restoration of the host’s immune system—either through antiretroviral therapy in HIV patients or the cessation of immunosuppressants. However, for many patients, these strategies are either insufficient or impossible. The emergence of immune checkpoint inhibitors (ICIs), such as those targeting the PD-1/PD-L1 axis, offered a glimmer of hope, theoretically ‘releasing the brakes’ on exhausted antiviral T-cells. Yet, clinical responses to ICIs have been notoriously heterogeneous. Two landmark studies published in JAMA Neurology by Möhn and colleagues now provide critical insights into why some patients respond while others do not, and how allogeneic cell therapy might fill the gap for the most vulnerable.
Endogenous Virus-Specific T-Cells: A Powerful Biomarker for ICI Response
The first study (Möhn et al., 2026) addressed a pivotal question: Can we predict which PML patients will benefit from immune checkpoint inhibition? This retrospective cohort study analyzed 111 patients across 39 centers, making it one of the largest investigations into ICI use for PML to date. The researchers focused on the presence of preexisting JCV- and/or BK virus-specific T-cells in the peripheral blood prior to the initiation of therapy.
Study Design and Patient Exposure
The cohort included 111 patients (median age 61 years) who received pembrolizumab, nivolumab, or atezolizumab. The primary exposure variable was the detection of virus-specific T-cells (VSTs) using ELISpot or flow cytometry before treatment. Patients were stratified into T-cell positive, T-cell negative, and unknown status groups. The median follow-up was 7 months, providing a robust window to observe clinical and virological outcomes.
The Prognostic Power of Preexisting Immunity
The results were striking. Patients who had detectable VSTs before starting an ICI had a clinical response rate of 86% (18/21). In sharp contrast, those who were T-cell negative had a response rate of only 23% (5/22). This difference translated into a significant survival advantage. While the median survival time for the T-cell positive group was not reached during the study period, the T-cell negative group faced a median survival of just 136.5 days. Furthermore, T-cell positive patients achieved superior functional outcomes, as measured by the modified Rankin Scale, and demonstrated a more rapid clearance of JC viral load from the cerebrospinal fluid (CSF).
Perhaps most surprisingly, the study found that immune-related adverse events (irAEs) were more frequent and severe in the T-cell negative group (50%) compared to the T-cell positive group (10%). This suggests that in the absence of a focused antiviral response, checkpoint inhibition may lead to dysregulated, off-target inflammation rather than productive immunity.
Directly Isolated Allogeneic Virus-Specific T-Cells: A New Hope for the Immunocompromised
While the first study established that ICIs require a ‘foundation’ of existing immunity to be effective, it left a critical void: What happens to the patients who are T-cell negative? The second study (Möhn et al., 2024) explored the use of directly isolated allogeneic virus-specific (DIAVIS) T-cells as a potential solution.
The DIAVIS Approach: Methodology
This retrospective case series included 28 patients with PML who received T-cells isolated from healthy donors. Because JC virus and BK virus share significant genomic homology, healthy donors with high titers of BK-specific T-cells were used. The T-cells were isolated within 24 hours and administered at a maximum dose of 2 x 10^4 CD3+ cells/kg. This adoptive transfer aimed to provide the patient with the cellular machinery necessary to combat the viral infection immediately.
Clinical Stabilization and Survival Comparison
The results for DIAVIS were highly encouraging. Of the 28 patients, 22 (79%) responded to the treatment, showing clinical stabilization or improvement and a reduction in CSF viral load. When compared to a historical reference group receiving best supportive treatment (BST), the DIAVIS group showed a significantly higher 12-month survival rate (69% vs 45%). The hazard ratio for death was 0.42, indicating a substantial protective effect. Older age was identified as the primary predictor of poor response, likely reflecting a more general decline in the recipient’s ability to support the transferred cells or underlying frailty.
Synthesizing the Evidence: A New Treatment Algorithm
Together, these two studies suggest a shift in how clinicians should approach PML. The presence of peripheral VSTs appears to be the ‘gatekeeper’ for therapeutic success.
1. For VST-positive patients: Immune checkpoint inhibitors like pembrolizumab are highly likely to be effective. These patients possess the specific T-cells required; they simply need the ICI to overcome viral-induced exhaustion.
2. For VST-negative patients: ICIs are likely to fail and may even cause harm through severe irAEs. In these cases, adoptive transfer of allogeneic VSTs (DIAVIS) should be considered as an early intervention to provide the missing immune response.
This precision-medicine approach could fundamentally change the standard of care, moving away from a ‘one-size-fits-all’ trial of ICIs toward a biomarker-driven strategy.
Expert Commentary and Methodological Considerations
These findings are practice-changing, but they must be interpreted within the context of their study designs. Both studies were retrospective, which introduces inherent biases in patient selection and data collection. The use of historical controls in the DIAVIS study, while necessary given the rarity and severity of the disease, is not as robust as a randomized controlled trial.
Furthermore, the logistical challenges of DIAVIS therapy cannot be overlooked. The need for rapid donor screening, T-cell isolation, and administration within a narrow window requires highly specialized centers. However, the high response rate (79%) in a population that typically faces a terminal diagnosis is a powerful argument for expanding access to these technologies.
From a mechanistic perspective, the data reinforce the importance of the T-cell receptor (TCR) repertoire. In PML, the JC virus often undergoes mutations in the VP1 capsid protein that allow it to evade immune detection. By providing a diverse array of VSTs from healthy donors, DIAVIS therapy may overcome this viral escape, providing a broader ‘net’ to catch and clear the infection.
Conclusion
The management of Progressive Multifocal Leukoencephalopathy is entering a new era. We now know that the efficacy of immune checkpoint inhibitors is deeply tethered to the patient’s underlying antiviral T-cell reservoir. By utilizing VST testing as a predictive biomarker, clinicians can better select candidates for ICI therapy. For those lacking this immunity, the adoptive transfer of allogeneic T-cells offers a life-saving alternative. Future prospective trials are urgently needed to codify these strategies into formal guidelines, but the path toward significantly improved survival in PML is now clearer than ever.
References
1. Möhn N, Grote-Levi L, Bonifacius A, et al. Virus-Specific T Cells and Response to Checkpoint Inhibitors in Progressive Multifocal Leukoencephalopathy. JAMA Neurol. 2026; doi:10.1001/jamaneurol.2025.5318.
2. Möhn N, Grote-Levi L, Wattjes MP, et al. Directly Isolated Allogeneic Virus-Specific T Cells in Progressive Multifocal Leukoencephalopathy. JAMA Neurol. 2024;81(11):1187–1198. doi:10.1001/jamaneurol.2024.3324.
3. Cortese I, Muranski P, Enose-Akahata Y, et al. Pembrolizumab Treatment for Progressive Multifocal Leukoencephalopathy. N Engl J Med. 2019;380(17):1597-1605.
