Progressive multifocal leukoencephalopathy (PML) is a viral encephalitis caused by the John Cunningham (JC) virus (Schweitzer et al., 2023). First described in 1958 in patients with hematologic malignancies, PML typically affects individuals with severe immunosuppression. This can result from acquired immune deficiency syndrome (AIDS), hematologic malignancies, solid organ transplantation, or immunosuppressive therapies (White and Khalili, 2011). JCV, an ubiquitous neurotropic polyomavirus, remains latent in the kidneys and bone marrow of healthy individuals (Tan and Koralnik, 2010). PML diagnosis relies on established neuropathological findings or combined clinical, radiological, and laboratory criteria (Berger et al., 2013; Cortese et al., 2021)( Figure 1 ). The disease is characterized by lytic infection of oligodendrocytes, leading to white matter demyelination, primarily in immunocompromised hosts (Gheuens et al., 2013). In human immunodeficiency virus (HIV)-infected patients receiving antiretroviral therapy (ART), immune reconstitution may trigger immune reconstitution inflammatory syndrome (IRIS) in some PML cases. Notably, evolving understanding of disease progression has revealed novel histological characteristics identified during routine diagnostic evaluation.

The objective of this study is to systematically assess the clinical and pathological characteristics of PML within the AIDS patient cohort. Enhancing the depth of knowledge regarding this disease is instrumental in refining diagnostic accuracy for pathologists and in paving the way for innovative approaches in the exploration of underlying mechanisms and the development of therapeutic interventions.

The JC virus is prevalent in the human body, particularly in the kidney and brain tissues. It seldom causes disease in individuals with normal immune systems; however, it can reactivate in those with compromised immunity (e.g., a CD4+ T lymphocyte count of less than 100 cells/uL) (Brew et al., 2010). HIV and JCV have a synergistic effect that can lead to PML. When AIDS is complicated by central nervous system infections, the blood-brain barrier’s permeability to JCV increases. The HIV Tat protein or chemokines secreted by it can stimulate the replication of JCV (Enam et al., 2004). The infection also affects cerebellar granule cell neurons and cortical pyramidal neurons within the gray matter, resulting in extensive and multi-stage demyelination. As the National Infectious Disease Center (Beijing), our center has treated a large number of HIV patients. As the number of diagnosed cases increases, we have observed numerous new findings in clinical biochemical parameters, imaging features, and pathomorphological characteristics that diverge from those described in traditional textbooks and literature.

Falcó et al. have shown that the occurrence of PML is related to lower blood CD4 levels (Falcó et al., 2008). In our study, we also found that over ninety percent of the cases had low blood CD4 levels, but close to forty percent of the cases had CD4 levels > 200 cells/μl. Furthermore, we also found that over eighty percent of the cases underwent biopsy or surgery after starting ART. This suggests that even in HIV patients with CD4 levels > 200 cells/μl and standardized ART, vigilance for the occurrence of PML is still necessary.

The study of Trunfio has shown that the typical MRI manifestations of PML are usually confined to the white matter, involving the subcortical “U” fibers, with a predilection for the parietal, occipital, and frontal lobes, although they can also occur in the basal ganglia, thalamus, and corpus callosum. Edema and mass effect are uncommon, and lesion enhancement is rare. The lesions are irregular in shape, with low signal intensity on T1WI, high signal intensity on T2WI/2-FIAIR, and no restricted diffusion on DWI, although there have been reports of peripheral enhancement, especially in treated HIV patients (Trunfio et al., 2019). Our study revealed similar imaging findings, with the difference being that approximately 20% of cases exhibited mass effect and about 30% exhibited enhancement. Furthermore, our study found that mass effect was more likely to occur in cases with higher blood CD4+ T-cell counts (P = 0.022) and higher CSF protein (P < 0.001), while enhancement was more likely to occur in cases with higher CSF protein (P = 0.032).

Due to the possibility that patients may be very weak and have a lower resistance, imaging examinations combined with JCV PCR of cerebrospinal fluid are typically used to diagnose PML. These patients avoid invasive brain biopsy. Whiteman showed that before the introduction of ART, PCR for JCV-DNA was 72%-92% sensitive and 92%-100% specific for the diagnosis of PML (Augusto et al., 2015; Whiteman et al., 1993). However, Cinque found that it has become common to have negative CSF molecular results in patients with AIDS with clinical and imaging presentations indistinguishable from those of PML. This may be due to the immune restoration with antiretroviral therapy leading to a decreased viral replication and increased clearance of JCV DNA from the CSF (Cinque et al., 2003). As a result of this, the sensitivity of PCR testing for JCV-DNA is now estimated to be around 58% (Antinori et al., 2003; Marzocchetti et al., 2005). In our group of 19 cases, 6 underwent cerebrospinal fluid NGS testing before surgery, and 50% (3/6) of them detected JCV. For cases where imaging or cerebrospinal fluid etiology cannot be clearly defined, brain biopsy becomes particularly important. These cases with atypical imaging or cerebrospinal fluid pathogens often exhibit more changes in histology.

The study of Berger et al. indicates that the histopathological features of PML are characterized by multifocal demyelinating lesions predominantly affecting the white matter, enlargement of oligodendrocyte nuclei, chromatin with a ground-glass appearance, and increased size of astrocytes with bizarre morphology and irregular, darkly stained nuclei. The inflammatory response is relatively mild, primarily consisting of lymphocytic infiltration, often seen around blood vessels. Neurons are relatively preserved, with lesions primarily affecting the white matter and glial cells, and late-stage tissue necrosis and cavitation may occur (Berger, 2014). All 19 cases in this study were confirmed to have JCV infection through molecular testing (10 by JCV-specific PCR and 17 by mNGS of lesion tissues). Among the 17 cases tested by mNGS, 4 exhibited co-infections. Given the consistently low relative abundance of co-infecting pathogens (0.01%-0.67%), we primarily attribute the pathological changes to JCV infection. Our study indicates that over 60% of cases exhibit moderate to severe inflammation, and nearly half of the cases show active inflammation. Our research also found that cases with higher blood CD8+ T-cell counts are more likely to experience severe inflammatory responses (P = 0.022), including a significant number of neutrophils (P = 0.001), plasma cell infiltration (P = 0.004), and lymphatic vessel cuffing (P = 0.012). Cases that have undergone ART were also more likely to experience severe inflammatory responses (P = 0.036). Cases with an HIV diagnosis duration exceeding one month were also more likely to experience severe inflammatory responses (P = 0.038), significant plasma cell infiltration (P = 0.011), and lymphatic vessel cuffing (P = 0.005). Cases with enhanced imaging findings were more likely to show moderate to severe inflammation (P = 0.044), have plasma cell infiltration (P = 0.011) and lymphatic vessel cuffing histologically, rather than gitter cells (P = 0.018). These cases display characteristics that deviate from the classic PML previously reported, exhibiting a tendency towards MRI enhancement and histologically indicating a more severe inflammatory response, especially for patients following ART treatment.

Many studies on the immune response against PML have focused on the role of T cells. Patients exhibiting stronger JCV-specific T cell responses tend to have a better prognosis (Khanna et al., 2009). The research of Khanna et al. Showed that survival was correlated to an increase in CD4+ but not in CD8+ JCV-specific T cells. Du and Koralnik et al. showed that within the CNS, however, CD8+ T cells appear to play a critical role. They also proved that the presence of JCV-specific cytotoxic T lymphocytes in the peripheral blood at the time of diagnosis is associated with a better clinical outcome (Du Pasquier et al., 2004; Koralnik et al., 2002; Lima et al., 2007).

Our research results also indicated that cases with higher blood CD8+ T-cell counts are more prone to immune reactions. Furthermore, among the 19 cases studied, 4 exhibited very significant inflammation, with 1 case clinically confirmed to have developed PML-IRIS. Immune reconstitution can be established in HIV patients treated with ART. Forced immune reconstitution frequently may result in profound clinical exacerbation of PML, often associated with high morbidity and mortality. During the immune reconstitution process, cytotoxic T cells eliminate a wide range of cells infected with JCV. Consequently, the excessive immune response exacerbates tissue injury within the lesions (Bauer et al., 2015).

Despite proposed diagnostic criteria (Müller et al., 2010)( Figure 1 ), IRIS manifestations are diverse and lack precise definition (Kaplan et al., 2009), making its diagnosis and the distinction between PML and PML-IRIS clinically challenging. Typically characterized by fever and deterioration of the underlying opportunistic infection, differentiating IRIS from disease progression due to new/worsening infection remains difficult yet critical due to the potential need for risky corticosteroids. Most IRIS cases present within 4–8 weeks of ART initiation in patients with high pre-ART viral loads and low CD4+ counts, though later onset or occurrence in sequestered sites is possible.

In our 12 cases with severe inflammation, the time for ART initiation ranged from 1 to 48 weeks, among which 4 cases with very severe inflammation, the time for ART treatment ranged from 8 to 12 weeks( Supplementary Table 3 ). Our clinically diagnosed case of PML-IRIS underwent 12 weeks of ART exhibited mass effect on imaging, accompanied by patchy enhancement, and histologically there was a significant infiltration of lymphocytes and plasma cells. After corticosteroid treatment, the general condition of this case improved, and the patient was discharged in stable condition. Jan et al. demonstrated that the primary difference between PML and PML-IRIS is the quantity of infiltrating immune T and B cells (Bauer et al., 2015). However, our research suggests that the infiltration of a large number of inflammatory cells is not exclusive to PML-IRIS.

In our cohort of 19 cases, 16 patients with PML received preoperative ART, with durations ranging from 7 days to 1 year. Among these, 4 cases presented with CNS infections as the initial manifestation. Of the remaining 12 cases, 6 exhibited worsening of clinical symptoms 1–3 months after ART initiation. Histopathological examinations revealed prominent inflammatory infiltrates in 4 cases, supporting a diagnosis of PML-IRIS, with one case confirmed clinically. In contrast, 2 cases demonstrated mild histological inflammation, which was insufficient for a PML-IRIS diagnosis, suggesting symptom exacerbation unrelated to immune reconstitution ( Supplementary Table 4 ). Our research indicates that PML and PML-IRIS constitute a continuous process with an intermediate link that may exist. Due to the complexity of clinical symptoms and imaging manifestations, neurologists often find it challenging to clearly define these cases, and may even suspect neoplastic lesions. This phenomenon was initially overlooked in our diagnosis, suggesting that clinicians consider other etiological infections. After the mNGS sequencing, it was officially confirmed as a simple JC virus infection.

In clinical practice, an increasing number of PML cases no longer require pathological biopsy to confirm the presence of PML. Pathologists are increasingly encountering so-called “atypical” cases, which also exhibit atypical tissue morphology. These cases present higher demands and challenges for accurate diagnosis by pathologists. Given the rarity of these cases, further data accumulation is necessary for this study. Additionally, there is a scarcity of explorations into the underlying mechanisms.

Our study reports on typical PML cases and also describes a unique category of PML cases characterized by the presence of enhancement and mass effect on MRI, in addition to the classical PML. Histologically, there is a more severe inflammatory response, particularly in patients undergoing ART, where the imaging and pathological morphology overlaps with that of PML-IRIS. The inflammatory response observed histologically occurs before the clinical onset of IRIS, and a definitive diagnosis can alert clinicians to the possibility of IRIS, prompting early intervention and treatment.