Contribution of cytotoxic CD8 T cells, neutrophils and type 1 interferon signaling to hyperinflammatory pathology in HIV associated TB meningitis

Using single-cell RNA sequencing of cerebrospinal fluid from HIV-associated tuberculous meningitis patients, this study identifies a hyperinflammatory pathology characterized by the accumulation of granzyme-rich cytotoxic CD8 T cells, highly activated neutrophils, and detrimental type 1 interferon signaling that correlates with bacterial load and treatment response.

The Gist

The Big Picture: A "Friendly Fire" Disaster in the Brain

Imagine your brain is a high-security fortress. When a dangerous invader (the bacteria that causes Tuberculosis, or TB) sneaks in through the door (the spinal fluid), the body's security team (the immune system) rushes to defend it.

In most people, this defense is organized and effective. But in people living with HIV who get TB meningitis (TB in the brain), the defense team goes haywire. Instead of just fighting the bacteria, they start destroying the fortress itself. This is called hyper-inflammation.

This paper is like a high-tech security camera feed (using a technology called single-cell RNA sequencing) that zoomed in on 188,000 individual immune cells in the spinal fluid of 25 patients. The researchers wanted to figure out: Who is causing the damage, and why?

Here are the three main "villains" they found:

1. The Overzealous "Grunt" Soldiers (Cytotoxic CD8 T Cells)

The Analogy: Imagine a squad of soldiers whose only job is to shoot anything that moves. In a normal battle, they aim only at the enemy. In these patients, the squad is huge, and they are carrying "grenades" (proteins called Granzymes) that they drop everywhere, even on their own side.

• What the paper found: The most common cell in the brain fluid was a type of CD8 T cell. These cells were packed with "granzymes" (specifically one called GZMK).
• The Twist: Usually, these soldiers are supposed to kill infected cells. But here, they were mostly just "standing around" with their grenades ready, not actually killing the TB bacteria effectively.
• The Damage: The paper discovered that the GZMK protein acts like a trigger for the body's "complement system" (a chemical chain reaction). It's like these soldiers are accidentally pulling the pin on a grenade that sets off a massive chain reaction, blowing up the brain tissue and causing swelling.
• The HIV Factor: Because HIV eats away at the "generals" (CD4 T cells), the "grunts" (CD8 cells) take over the whole operation, leading to a chaotic, uncoordinated defense.

2. The "Recruitment Siren" (Neutrophils)

The Analogy: Think of neutrophils as the first responders—firefighters who rush to a fire. Usually, they put out the fire and leave. In these patients, the firefighters are stuck in a loop. They are screaming "Help! More firefighters needed!" (releasing a chemical called IL-8) even though the fire is already out of control.

• What the paper found: The neutrophils in the brain were highly activated. They weren't just fighting; they were expressing genes that told the body to send more neutrophils from the blood.
• The Damage: This creates a "feed-forward" loop. More neutrophils arrive, they release more "sirens," and even more arrive. This pile-up of cells causes massive pressure and damage to the delicate brain tissue.
• The Link to Bacteria: The more bacteria the patient had, the more "sirens" were being screamed, and the more chaotic the scene became.

3. The "False Alarm" Signal (Type 1 Interferon)

The Analogy: Imagine a smoke detector that is so sensitive it goes off when you just toast bread. This signal tells the body, "We are under viral attack! Lock down everything!" But in TB, this signal is actually counter-productive. It shuts down the "special forces" (the helpful immune cells) and tells the "chaos squad" (the inflammatory cells) to keep going.

• What the paper found: The patients had high levels of Type 1 Interferon signaling. This is a signal usually meant for viral infections (like the flu), not bacterial ones.
• The Damage: This signal suppresses the body's ability to fight the TB bacteria effectively. It's like the general telling the army to stop shooting and start hiding, while the enemy (TB) keeps multiplying.
• The Surprise: Even after the patients started taking antibiotics to kill the bacteria, this "False Alarm" signal didn't stop. In fact, it got louder in the blood and brain fluid over the next few weeks. This explains why patients often stay sick and inflamed for a long time, even after the bacteria are dying.

The "Smoking Gun": Bacterial Load Matters

The researchers noticed a clear pattern:

• Low Bacteria: The immune response was messy, but manageable.
• High Bacteria: The immune system went into "Total War" mode. The "grunts" (CD8 cells) got more aggressive, the "firefighters" (neutrophils) screamed louder, and the "false alarm" (Interferon) blared continuously.

Why Does This Matter?

For a long time, doctors have tried to treat this condition with steroids (to calm the immune system), but it hasn't worked well, especially for people with HIV.

This paper suggests that the current "calm down" approach might be too broad. Instead, we might need targeted therapies that:

1. Stop the "grunts" from triggering the complement chain reaction (blocking GZMK).
2. Turn off the "recruitment siren" so no more neutrophils rush in (blocking IL-8 or CXCR2).
3. Silence the "False Alarm" (Type 1 Interferon) so the helpful immune cells can do their job.

The Bottom Line

In HIV-associated TB meningitis, the body's immune system isn't just fighting the bacteria; it's accidentally burning down the house it's trying to protect. The bacteria act as the spark, but the immune system's overreaction (driven by specific cells and signals) is the fire. To save lives, future treatments need to stop the fire, not just the spark.

Technical Summary

1. Problem Statement

Tuberculous meningitis (TBM) is the most severe form of tuberculosis (TB), carrying high mortality and morbidity rates, particularly in people living with HIV (PLWHA). Despite the high burden of disease, adjunctive anti-inflammatory therapies (such as corticosteroids) have failed to improve survival in PLWHA. Current understanding of the Central Nervous System (CNS) immune response in HIV-associated TBM is limited, hindering the development of host-directed therapies. A critical gap exists in understanding how bacterial load correlates with specific immune cell phenotypes and transcriptomic signatures in the cerebrospinal fluid (CSF) of these patients.

2. Methodology

The study employed a multi-modal, high-resolution approach combining single-cell and bulk RNA sequencing across two distinct cohorts:

• Cohort 1 (Single-Cell RNA Sequencing - scRNAseq):
  • Samples: 25 lumbar CSF samples from adults with HIV-associated TBM (enrolled in the INTENSE-TBM trial).
  • Technology: 10x Genomics Fixed RNA Profiling (probe-based) on fixed cells collected 7 days post-enrollment.
  • Scale: 188,983 cells were analyzed after quality control (ambient RNA and doublet removal).
  • Analysis: Cells were clustered and annotated using canonical markers and differential gene expression (DEG). Pseudobulk analysis was performed to compare microbiologically confirmed (Mtb+) vs. unconfirmed (Mtb-) cases, adjusting for age and sex.
• Cohort 2 (Longitudinal Bulk RNA Sequencing):
  • Samples: CSF and whole blood from a separate cohort (LASER-TBM and INTENSE-TBM trials).
  • Timepoints: CSF at Day 7 and Week 4; Blood at Day 0, Day 7, Week 2, and Week 8.
  • Analysis: Bulk RNA-seq to track temporal changes in interferon signaling and inflammatory pathways during antibiotic treatment.
• Bioinformatics: Utilized Seurat, sccomp (for differential composition), DESeq2 (for DEG), and GSEA (Gene Set Enrichment Analysis).

3. Key Contributions

• Largest scRNAseq Dataset in TBM: This is the largest study to date characterizing the CSF immune landscape in HIV-associated TBM, providing unprecedented resolution of cell heterogeneity.
• Identification of a Hyper-inflammatory Signature: The study defines a specific "hyper-inflammatory" state in the CSF characterized by cytotoxicity, neutrophil activation, and detrimental Type 1 Interferon (IFN) signaling.
• Bacterial Load Correlates: It establishes distinct transcriptomic correlates between bacterial load (microbiological confirmation) and specific immune cell phenotypes, suggesting that bacterial burden drives specific immunopathological mechanisms.
• Temporal Dynamics: It reveals that Type 1 IFN signaling paradoxically increases after the initiation of antibiotic therapy, contrasting with the resolution seen in pulmonary TB.

4. Key Results

A. Cellular Composition and Heterogeneity

• CD8 T Cell Predominance: Contrary to previous studies where CD4 cells dominate, CD8 T cells were the predominant cell type in the CSF (mean 63.1%), even in patients with CD4 counts >500 cells/mm³. This is likely due to HIV-mediated depletion of CCR5+/CXCR4+ CD4 cells.
• Cytotoxic Phenotype: The majority of CD8 T cells exhibited a cytotoxic effector memory phenotype with high expression of granzymes (GZMA, GZMM, GZMK) and perforin, but low cytokine expression (e.g., IFNG).
• Neutrophil Activation: Neutrophils were highly activated, expressing markers of enhanced effector function (ICAM1, NAMPT) and split into two phenotypes: one expressing MMP25/IL1B/SOCS3 and another expressing high IL-8 (CXCL8) and CXCR2, suggesting a "feed-forward" loop for neutrophil recruitment.

B. Impact of Bacterial Load (Mtb+ vs. Mtb-)

• Increased Cytotoxicity: Patients with microbiologically confirmed TBM (higher bacterial load) showed significantly increased expression of cytotoxic genes (GZMB, GZMH, PRF1, NKG7) in CD4 T cells, MAIT cells, and NK cells.
• Type 1 IFN Signaling: There was a marked upregulation of Type 1 Interferon Stimulated Genes (ISGs) in T cells, B cells, NK cells, and dendritic cells in Mtb+ patients.
• Specific Clusters:
  • CD8 T Cells: Increased abundance of GZMK+ and GNLY+ clusters in Mtb+ cases.
  • CD4 T Cells: An IFNG+ cluster in CD4 cells showed a ~25-fold increase in GZMB expression in Mtb+ patients, suggesting a shift from polyfunctional cytokine production to cytotoxicity under high bacterial load.

C. Temporal Dynamics of Type 1 IFN

• Delayed and Sustained Response: Unlike pulmonary TB where IFN signaling resolves with treatment, Type 1 IFN signaling in HIV-associated TBM increased between Day 7 and Week 4 in the CSF and remained elevated in the blood up to Week 8.
• Pathway Enrichment: Gene Set Enrichment Analysis (GSEA) confirmed that Type 1 IFN pathways were the top upregulated pathways at Week 4 in the CSF.
• Mechanism: This sustained signaling may explain the persistence of CNS inflammation and neutrophil recruitment despite antibiotic therapy.

5. Significance and Implications

• Immunopathology Mechanism: The study proposes that the poor outcomes in HIV-TBM are driven by a "hyper-inflammatory" loop:
  1. High bacterial loads drive cytotoxicity (via GZMK, GZMB) and Type 1 IFN signaling.
  2. GZMK cleaves complement proteins (C2/C4), activating the complement cascade and promoting inflammation.
  3. Type 1 IFN suppresses protective Th1 responses and promotes neutrophil recruitment (via IL-8), creating a "feed-forward" cycle of tissue damage.
• Therapeutic Targets:
  • Type 1 IFN Blockade: Given the detrimental role of sustained Type 1 IFN, blocking this pathway (e.g., via anti-IFNAR antibodies) could be a viable host-directed therapy.
  • Neutrophil Inhibition: Targeting neutrophil chemotaxis (e.g., CXCR2 inhibitors) might break the cycle of intracranial inflammation.
  • Complement Inhibition: The role of GZMK in activating complement suggests complement inhibitors could mitigate immunopathology.
• Stratification: The findings suggest that future clinical trials for host-directed therapies must stratify patients based on bacterial load and specific immune phenotypes (e.g., neutrophil count, IFN signature) rather than treating all TBM patients as a homogeneous group.

Limitations

• Probe-Based Sequencing: The 10x Flex platform used did not capture variable regions of T-cell receptors (TCR) or B-cell receptors (BCR), limiting clonality analysis.
• Selection Bias: Patients with very low CSF cell counts were excluded, potentially biasing the cohort toward more severe inflammation.
• Corticosteroids: All patients received corticosteroids, which may have modulated the observed gene expression profiles.
• Lack of Healthy Controls: Due to the invasiveness of lumbar puncture, healthy controls were not included; comparisons were made against non-TBM meningitis or historical data.

In conclusion, this study provides a high-resolution map of the HIV-TBM immune landscape, identifying a distinct, detrimental hyper-inflammatory state driven by cytotoxic CD8 T cells, activated neutrophils, and sustained Type 1 IFN signaling, offering new targets for improving survival in this high-risk population.