Human neurons stimulated with IFNγ present HLA class I-restricted autoantigens to cytotoxic CD8+ T cells

This study demonstrates that IFNγ stimulation induces human neurons to present a distinct, HLA-B-enriched repertoire of autoantigens via HLA class I molecules, which can be recognized by autoreactive CD8+ T cells to cause antigen-specific neuronal injury, thereby providing a platform for discovering and validating neuronal autoantigens in inflammatory CNS disorders.

The Gist

Imagine your brain is a bustling city, and the neurons are the hardworking citizens keeping everything running. Usually, these neurons are like private citizens; they don't wear "ID badges" (called HLA class I molecules) that show their internal contents to the city's security guards (the immune system's CD8+ T cells). This keeps them safe from being attacked by mistake.

However, the paper describes what happens when the city gets into a state of high alert, triggered by a signal called IFN-gamma (a chemical messenger often found during inflammation in neurological disorders).

Here is the story of what the researchers discovered, broken down into simple steps:

1. Putting on the ID Badges

When the neurons receive this "high alert" signal (IFN-gamma), they suddenly start wearing ID badges. Before this, they were invisible to the security guards. Now, they are displaying a list of their own internal proteins on the surface, just like a citizen holding up a menu of what they are made of.

2. The "Menu" of the Neuron

The researchers used a special high-tech microscope (mass spectrometry) to read these menus. They found that when neurons are stressed by IFN-gamma, they display a specific set of tiny protein fragments (peptides).

• The Analogy: Think of it like a restaurant that suddenly changes its menu. The researchers found that the "neuron menu" is very specific, mostly featuring 9-letter "words" (peptides) made from proteins that only neurons have.

3. The Security Guard's Reaction

To test if this actually causes trouble, the scientists created a "fake" neuron menu using a special genetic switch. They put a known "target" on the neurons that only appeared when the IFN-gamma signal was active.

• The Result: When the security guards (CD8+ T cells) saw this specific target on the neuron's ID badge, they got angry. They recognized it as something to attack and proceeded to damage the neuron's long arms (neurites), effectively injuring the cell. This proves that if the immune system sees these specific neuron parts, it can attack the brain tissue.

4. The "Specialist" Guards (HLA-B)

The researchers compared the menus of neurons against other body cells (like skin cells) and found that the neuron menus were unique.

• The Discovery: They noticed that the "specialist guards" (specifically those with HLA-B badges) were the ones most likely to pick up and display these neuron-specific targets. It's as if the neuron's stress signals are perfectly tuned to be seen by this specific type of security guard.

5. Proving the Source

To make sure they were actually seeing the neurons' own proteins and not just random noise, the scientists did a clever trick. They removed the "base" of the ID badge system (beta-2-microglobulin) from the neurons.

• The Result: The menus disappeared. When they put the base back in, the menus reappeared, including a specific target called NEFL (a part of the neuron's skeleton). This target was found across different people, suggesting it's a common "weak spot" that the immune system might recognize in many humans.

The Big Picture

In short, this paper builds a new laboratory "playground" where they can watch neurons in a bottle. They found that when the brain is inflamed (IFN-gamma), neurons start wearing ID badges that show their own internal parts. This creates a landscape where specific immune guards (especially HLA-B types) can spot these neurons, recognize them as targets, and attack them.

This explains a potential mechanism for how the immune system might accidentally turn against the brain during inflammatory neurological diseases: the brain cells themselves, under stress, put up a "target on my back" sign that the immune system can read and act upon.

Technical Summary

Technical Summary

Problem Statement
Interferon-gamma (IFN$\gamma$) signaling is a hallmark of inflammatory environments in the central nervous system (CNS) across various neurological disorders. However, the specific neuronal peptides presented on Major Histocompatibility Complex (MHC) class I molecules under these inflammatory conditions, and the subsequent functional consequences for cytotoxic CD8+ T cells, remain incompletely defined. A critical gap exists in understanding how IFN$\gamma$ alters the neuronal immunopeptidome and whether these changes enable the recognition and destruction of neurons by autoreactive T cells.

Methodology
To address this, the authors developed an integrated platform combining three primary technical approaches:

1. Human iPSC-Derived Neural Aggregates (HNAs): The study utilized HNAs as a physiologically relevant human neuronal model.
2. Immunopeptidomics: The team employed HLA class I immunoprecipitation coupled with liquid chromatography-tandem mass spectrometry (LC-MS/MS) to map the peptide ligands presented on the cell surface.
3. Microfluidic Co-culture Assays: These assays were used to test antigen-specific cytotoxicity in a controlled environment.

The experimental design involved stimulating HNAs with IFN$\gamma$ to induce HLA class I expression. To validate antigen presentation and T cell reactivity, the authors utilized a synapsin-driven polyepitope cassette to express defined neoantigen 9-mer peptides in a neuron-restricted manner. Comparative immunopeptidomics were performed across four donors, contrasting IFN$\gamma$-treated neurons with matched fibroblasts. Additionally, the study utilized $\beta$2-microglobulin ($\beta$2m) deletion and reconstitution strategies to confirm the neuronal origin of recovered peptides.

Key Results

• Induction of HLA Class I: IFN$\gamma$ stimulation induced robust HLA class I expression in HNAs, enabling the recovery of a canonical 8–12-mer ligandome enriched for 9-mers.
• Functional Cytotoxicity: Neuron-restricted expression of the polyepitope cassette resulted in the presentation of defined neoantigens. In the presence of IFN$\gamma$, this presentation elicited the activation of autologous, antigen-specific CD8+ T cells, leading to antigen-dependent neurite injury.
• Distinct Immunopeptidome: Comparative analysis revealed large repertoires of IFN$\gamma$-unique neural peptides that were distinct from those found on matched fibroblasts.
• HLA-B Enrichment: There was a consistent enrichment of predicted high-affinity binders on HLA-B allotypes within the neuronal peptide repertoire.
• Validation of Neuronal Origin: Deletion of $\beta$2-microglobulin ablated peptide recovery, while neuron-restricted reconstitution of $\beta$2m enabled the identification of neuron-derived peptides. Notably, peptides derived from neurofilament light (NEFL) were identified as being shared across donors and presented on multiple HLA allotypes.

Significance and Claims
The paper claims to provide an integrated platform for the discovery of neuronal autoantigens and their functional validation. The data support a specific model wherein IFN$\gamma$-driven neuronal HLA class I presentation creates an HLA-B-weighted epitope landscape. This landscape is capable of being recognized by autoreactive cytotoxic CD8+ T cells, offering a mechanistic explanation for how inflammatory signals in the CNS may facilitate T cell-mediated neuronal injury. The study establishes a direct link between inflammatory cytokine signaling, the alteration of the neuronal immunopeptidome, and the subsequent functional consequences for CD8+ T cell cytotoxicity.