Anti-PD-1 cancer immunotherapy induces central nervous system immune-related adverse events by microglia activation.

Cancer treatment with anti-PD-1 immunotherapy can cause central nervous system immune-related adverse events (CNS-irAEs). The role of microglia in anti-PD-1 immunotherapy-induced CNS-irAEs is unclear. We found that anti-PD-1 treatment of mice caused morphological signs of activation and major histocompatibility complex (MHC) class II up-regulation on microglia. Functionally, anti-PD-1 treatment induced neurocognitive deficits in mice, independent of T cells, B cells, and natural killer cells. Instead, we found that microglia mediated these CNS-irAEs. Single-cell RNA sequencing revealed major transcriptional changes in microglia upon anti-PD-1 treatment. The anti-PD-1 effects were mediated by anti-PD-1 antibodies interacting directly with microglia and were not secondary to peripheral T cell activation. Using a proteomics approach, we identified spleen tyrosine kinase (Syk) as a potential target in activated microglia upon anti-PD-1 treatment. Syk inhibition reduced microglia activation and improved neurocognitive function without impairing anti-melanoma effects. Moreover, we analyzed CNS tissue from a patient cohort that had received anti-PD-1 treatment. Imaging mass cytometry revealed that anti-PD-1 treatment of patients was associated with increased surface marker expression indicative of microglia activation. In summary, we identified a disease-promoting role for microglia in CNS-irAEs driven by Syk and provide an inhibitor-based approach to interfere with this complication after anti-PD-1 immunotherapy.

Targeting TGFß-activated kinase-1 activation in microglia reduces CAR T immune effector cell-associated neurotoxicity syndrome.

Cancer immunotherapy with chimeric antigen receptor (CAR) T cells can cause immune effector cell-associated neurotoxicity syndrome (ICANS). However, the molecular mechanisms leading to ICANS are not well understood. Here we examined the role of microglia using mouse models and cohorts of individuals with ICANS. CD19-directed CAR (CAR19) T cell transfer in B cell lymphoma-bearing mice caused microglia activation and neurocognitive deficits. The TGFß-activated kinase-1 (TAK1)-NF-κB-p38 MAPK pathway was activated in microglia after CAR19 T cell transfer. Pharmacological TAK1 inhibition or genetic Tak1 deletion in microglia using Cx3cr1CreER:Tak1fl/fl mice resulted in reduced microglia activation and improved neurocognitive activity. TAK1 inhibition allowed for potent CAR19-induced antilymphoma effects. Individuals with ICANS exhibited microglia activation in vivo when studied by translocator protein positron emission tomography, and imaging mass cytometry revealed a shift from resting to activated microglia. In summary, we prove a role for microglia in ICANS pathophysiology, identify the TAK1-NF-κB-p38 MAPK axis as a pathogenic signaling pathway and provide a rationale to test TAK1 inhibition in a clinical trial for ICANS prevention after CAR19 T cell-based cancer immunotherapy.

CSF1R inhibition promotes neuroinflammation and behavioral deficits during graft-versus-host disease in mice.

ABSTRACT: Chronic graft-versus-host disease (cGVHD) remains a significant complication of allogeneic hematopoietic stem cell transplantation. Central nervous system (CNS) involvement is becoming increasingly recognized, in which brain-infiltrating donor major histocompatibility complex (MHC) class II+ bone marrow-derived macrophages (BMDM) drive pathology. BMDM are also mediators of cutaneous and pulmonary cGVHD, and clinical trials assessing the efficacy of antibody blockade of colony-stimulating factor 1 receptor (CSF1R) to deplete macrophages are promising. We hypothesized that CSF1R antibody blockade may also be a useful strategy to prevent/treat CNS cGVHD. Increased blood-brain barrier permeability during acute GVHD (aGVHD) facilitated CNS antibody access and microglia depletion by anti-CSF1R treatment. However, CSF1R blockade early after transplant unexpectedly exacerbated aGVHD neuroinflammation. In established cGVHD, vascular changes and anti-CSF1R efficacy were more limited. Anti-CSF1R-treated mice retained donor BMDM, activated microglia, CD8+ and CD4+ T cells, and local cytokine expression in the brain. These findings were recapitulated in GVHD recipients, in which CSF1R was conditionally depleted in donor CX3CR1+ BMDM. Notably, inhibition of CSF1R signaling after transplant failed to reverse GVHD-induced behavioral changes. Moreover, we observed aberrant behavior in non-GVHD control recipients administered anti-CSF1R blocking antibody and naïve mice lacking CSF1R in CX3CR1+ cells, revealing a novel role for homeostatic microglia and indicating that ongoing clinical trials of CSF1R inhibition should assess neurological adverse events in patients. In contrast, transfer of Ifngr-/- grafts could reduce MHC class II+ BMDM infiltration, resulting in improved neurocognitive function. Our findings highlight unexpected neurological immune toxicity during CSF1R blockade and provide alternative targets for the treatment of cGVHD within the CNS.

Donor bone marrow-derived macrophage MHC II drives neuroinflammation and altered behavior during chronic GVHD in mice.

Graft-versus-host disease (GVHD) remains the leading cause of nonrelapse mortality after allogeneic stem cell transplantation for hematological malignancies. Manifestations of GVHD in the central nervous system (CNS) present as neurocognitive dysfunction in up to 60% of patients; however, the mechanisms driving chronic GVHD (cGVHD) in the CNS are yet to be elucidated. Our studies of murine cGVHD revealed behavioral deficits associated with broad neuroinflammation and persistent Ifng upregulation. By flow cytometry, we observed a proportional shift in the donor-derived T-cell population in the cGVHD brain from early CD8 dominance to later CD4 sequestration. RNA sequencing of the hippocampus identified perturbations to structural and functional synapse-related gene expression, together with the upregulation of genes associated with interferon-γ responses and antigen presentation. Neuroinflammation in the cortex of mice and humans during acute GVHD was recently shown to be mediated by resident microglia-derived tumor necrosis factor. In contrast, infiltration of proinflammatory major histocompatibility complex (MHC) class II+ donor bone marrow (BM)-derived macrophages (BMDMs) was identified as a distinguishing feature of CNS cGVHD. Donor BMDMs, which composed up to 50% of the CNS myeloid population, exhibited a transcriptional signature distinct from resident microglia. Recipients of MHC class II knockout BM grafts exhibited attenuated neuroinflammation and behavior comparable to controls, suggestive of a critical role of donor BMDM MHC class II expression in CNS cGVHD. Our identification of disease mediators distinct from those in the acute phase indicates the necessity to pursue alternative therapeutic targets for late-stage neurological manifestations.

Cep55 regulation of PI3K/Akt signaling is required for neocortical development and ciliogenesis.

Homozygous nonsense mutations in CEP55 are associated with several congenital malformations that lead to perinatal lethality suggesting that it plays a critical role in regulation of embryonic development. CEP55 has previously been studied as a crucial regulator of cytokinesis, predominantly in transformed cells, and its dysregulation is linked to carcinogenesis. However, its molecular functions during embryonic development in mammals require further investigation. We have generated a Cep55 knockout (Cep55-/-) mouse model which demonstrated preweaning lethality associated with a wide range of neural defects. Focusing our analysis on the neocortex, we show that Cep55-/- embryos exhibited depleted neural stem/progenitor cells in the ventricular zone as a result of significantly increased cellular apoptosis. Mechanistically, we demonstrated that Cep55-loss downregulates the pGsk3ß/ß-Catenin/Myc axis in an Akt-dependent manner. The elevated apoptosis of neural stem/progenitors was recapitulated using Cep55-deficient human cerebral organoids and we could rescue the phenotype by inhibiting active Gsk3ß. Additionally, we show that Cep55-loss leads to a significant reduction of ciliated cells, highlighting a novel role in regulating ciliogenesis. Collectively, our findings demonstrate a critical role of Cep55 during brain development and provide mechanistic insights that may have important implications for genetic syndromes associated with Cep55-loss.

Targeting BRF2 in Cancer Using Repurposed Drugs.

The overexpression of BRF2, a selective subunit of RNA polymerase III, has been shown to be crucial in the development of several types of cancers, including breast cancer and lung squamous cell carcinoma. Predominantly, BRF2 acts as a central redox-sensing transcription factor (TF) and is involved in rescuing oxidative stress (OS)-induced apoptosis. Here, we showed a novel link between BRF2 and the DNA damage response. Due to the lack of BRF2-specific inhibitors, through virtual screening and molecular dynamics simulation, we identified potential drug candidates that interfere with BRF2-TATA-binding Protein (TBP)-DNA complex interactions based on binding energy, intermolecular, and torsional energy parameters. We experimentally tested bexarotene as a potential BRF2 inhibitor. We found that bexarotene (Bex) treatment resulted in a dramatic decline in oxidative stress and Tert-butylhydroquinone (tBHQ)-induced levels of BRF2 and consequently led to a decrease in the cellular proliferation of cancer cells which may in part be due to the drug pretreatment-induced reduction of ROS generated by the oxidizing agent. Our data thus provide the first experimental evidence that BRF2 is a novel player in the DNA damage response pathway and that bexarotene can be used as a potential inhibitor to treat cancers with the specific elevation of oxidative stress.

Can computers conceive the complexity of cancer to cure it? Using artificial intelligence technology in cancer modelling and drug discovery.

Drug discovery and the development of safe and effective therapeutics is an intricate procedure, further complicated in the context of cancer research by the inherent heterogeneity and complexity of the disease. To address the difficulties of identifying, validating, and pursuing a promising drug target, artificial intelligence (AI) technologies including machine learning (ML) have been adopted at all stages throughout the drug development pipeline. Various methods are widely employed to efficiently process and learn from experimental data sets, with agent-based models garnering thorough interest due to their ability to model individual cell populations with aberrant phenotypes. The predictive power of artificial intelligence modelling techniques founded in comprehensive datasets and automated decision-making generates an obvious avenue of interest for application in the drug discovery pipeline.