🧬 biology

Granzyme A-expressing Terminal Effector T Cells Dedifferentiate into Long-lived Memory T Cells

This study challenges the strict bifurcation model by demonstrating that while early memory T cells primarily arise from Granzyme A-negative precursors, Granzyme A-expressing terminal effector T cells can subsequently dedifferentiate into long-lived functional memory cells, resulting in a mixed memory population derived from both linear and bifurcated differentiation paths.

Original authors: Nu Zhang, Chaoyu Ma, Wei Liao, Saranya Srinivasan, Chenhui he, Kenneth Fan, Tyng-An Zhou, Courtney Segura-Cepero, Heetanshi Jain, Sahana Jayakumar, Weiguo Cui, Yu Luan

Published 2026-06-26

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Nu Zhang, Chaoyu Ma, Wei Liao, Saranya Srinivasan, Chenhui he, Kenneth Fan, Tyng-An Zhou, Courtney Segura-Cepero, Heetanshi Jain, Sahana Jayakumar, Weiguo Cui, Yu Luan

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The Big Picture: How the Body Remembers Infections

Imagine your immune system is a massive army defending a castle (your body) against invaders (viruses). When a new enemy attacks, the army sends out a specialized squad of soldiers called CD8+ T cells.

For a long time, scientists thought these soldiers followed a strict, one-way street:

Training: They start as raw recruits.
Battle: They turn into fierce "Effector" soldiers who fight hard and kill the virus.
Retirement: Once the battle is won, most die off. A few survivors "retire" to become Memory T cells, sitting on the sidelines ready to fight again if the same enemy returns.

The big question was: Do the fierce fighters on the front lines ever get to become the wise veterans on the sidelines? Or do they have to choose one path or the other right at the start?

This paper says: Both things happen.

The New Discovery: The "Granzyme A" Badge

The researchers discovered a specific marker on these soldiers called Granzyme A. Think of Granzyme A as a heavy, red "KILLER" badge that some soldiers wear.

Granzyme B is like a standard uniform worn by almost all soldiers once the battle starts.
Granzyme A is a special, heavy-duty badge worn only by the most aggressive, "terminal" killers. These soldiers are so focused on destroying the enemy that they stop doing other things, like making cytokines (chemical signals that help coordinate the team).

The team created a special line of mice (let's call them the "Red Badge Mice") where any soldier who ever wore that heavy red Granzyme A badge would glow red forever, even if they took the badge off later. This allowed the scientists to track exactly who had been a "super-killer" in the past.

The Two Paths of the Soldiers

Using these glowing mice, the researchers watched what happened during a viral infection (using a virus called LCMV). They found two different stories playing out at the same time:

1. The "Early Choice" Path (The Bifurcation Model)

Right after the battle begins, the army splits into two groups:

Group A (The Future Veterans): These soldiers never put on the heavy red Granzyme A badge. They stay flexible, keep their "memory" potential, and quickly become the long-term Memory T cells.
Group B (The Terminal Killers): These soldiers put on the heavy red badge. They become super-aggressive killers but lose their ability to make chemical signals.

The Finding: At the very beginning of the memory phase (about a month after the infection), almost all the Memory T cells came from Group A. The "Terminal Killers" (Group B) seemed to be dead ends. This supports the idea that you have to choose your path early.

2. The "Late Comeback" Path (The Linear Model)

Here is the surprise. The researchers waited longer. They checked the army 4 to 6 months after the infection.

They found that the "Terminal Killers" (the ones who wore the red badge and then took it off) did not die out. Instead, they slowly changed back.

They shed their "killer" intensity.
They regained their ability to make chemical signals.
They transformed into fully functional Memory T cells.

The Finding: By the time the memory phase was mature (long-term), the pool of Memory T cells was a mix. It contained the "Early Veterans" (who never wore the badge) and the "Reformed Killers" (who wore the badge, fought hard, and then changed their minds).

What Does This Mean for the Army?

The paper concludes that the immune system uses a hybrid strategy:

Early on: It relies mostly on the soldiers who didn't go all-out killer mode. These are the ones who become the first line of long-term memory.
Later on: The soldiers who did go all-out killer mode eventually "dedifferentiate" (unlearn their killer habits) and join the memory ranks.

By the time the body is fully protected for the long haul, the "Memory Army" is a diverse team. It has the wise veterans who stayed calm, and the reformed warriors who once fought with extreme intensity. Both groups work together to protect the body.

A Simple Analogy: The Firefighters

Imagine a fire breaks out in a city.

The "Early Choice" Model: Some firefighters decide immediately to stay in the command center to learn the ropes (Memory). Others rush into the burning building with heavy hoses (Terminal Effectors). Traditionally, we thought the ones with the heavy hoses would burn out or die, and only the command center staff would remain.
This Paper's Discovery: The researchers found that the firefighters who rushed into the fire did survive. After the fire was out, they took off their heavy gear, cooled down, and eventually joined the command center as experienced veterans.

So, the final "Fire Department" (Memory T cells) is made up of both the people who stayed calm from the start and the people who fought the fire hard and then learned to be calm.

Summary of Key Claims

Granzyme A marks the most aggressive, short-term killer cells.
These aggressive cells lose their ability to make other signals (like IL-2) while they are fighting.
However, they are not dead ends. They can slowly change back into long-term Memory T cells.
The final pool of Memory T cells is a mixture of cells that chose the memory path early and cells that fought hard first and changed paths later.
Both paths are necessary to provide balanced, long-term protection.

Technical Summary: Granzyme A-expressing Terminal Effector T Cells Dedifferentiate into Long-lived Memory T Cells

Problem Statement
The differentiation trajectory of memory T cells ( $T_{MEM}$ ) remains a subject of debate, primarily characterized by two competing models: the linear model, where effector T cells lose effector functions to dedifferentiate into memory cells, and the bifurcation model, where early after priming, T cells diverge into distinct effector and memory precursor lineages. While recent evidence favors the bifurcation model—suggesting that terminal effector T cells lose the capacity to become memory cells—long-term in vivo tracing of terminal effector populations has been lacking. Specifically, it remains unresolved whether terminal effector T cells can contribute significantly to the long-term memory pool or if they represent a fixed, dead-end lineage. Furthermore, the distinct expression patterns and functional roles of granzyme A (Gzma) versus granzyme B (Gzmb) in CD8+ T cell differentiation are not fully understood.

Methodology
To address these gaps, the authors employed a combination of genetic fate mapping, adoptive transfer models, and acute viral infection systems:

Generation of Gzma-Fate Mapping (Gzma-FM) Mice: The authors generated a mouse strain by knocking a Cre recombinase cassette (Cre-T2A-BFP-P2A) into the endogenous Gzma locus, crossed with a Rosa26-LSL-tdTomato reporter. Due to low peptide skipping efficiency, the Gzma-Cre allele functions as a knockout, but the Cre activity permanently labels cells that have expressed Gzma (tdTomato+).
Viral Infection Models: Using Gzma-FM mice congenically marked with P14 TCR transgenes (specific for LCMV GP33-41), the authors infected mice with Lymphocytic Choriomeningitis Virus (LCMV) Armstrong. This allowed for the tracking of antigen-specific CD8+ T cells from the naive stage through the effector peak (days 7–8) to the memory phase (>30 days).
Comparative Analysis: The study compared Gzma-FM+ (cells with a history of Gzma expression) and Gzma-FM- (cells without) populations across multiple time points and tissues (spleen, lymph nodes, small intestine).
Functional Assays: The researchers assessed polyfunctionality (IFN- $\gamma$ , TNF- $\alpha$ , IL-2 production), surface marker expression (KLRG1, CX3CR1, IL-7R, CD62L, Tcf-1, T-bet), and recall responses. Recall experiments involved sorting memory cells (>60 days post-infection) and transferring them into naive recipients for secondary LCMV challenge.
Controls: A parallel Gzmb-FM mouse line was generated to demonstrate that Gzmb expression marks a broader, earlier population of activated cells, contrasting with the specific subset marked by Gzma.

Key Results

Distinct Expression Patterns: Unlike Gzmb, which is rapidly induced upon T cell activation, Gzma expression is distinct. In memory cells, Gzma protein levels decrease significantly upon TCR stimulation (likely due to rapid release from pre-formed stores), whereas Gzmb and cytokines are synthesized de novo.
Gzma Marks Terminal Effectors: At the peak of the immune response (day 7–8), the Gzma-FM+ population is enriched for terminal effector markers (high KLRG1, CX3CR1, T-bet; low IL-7R, CD62L) and lacks memory precursors (Tcf-1+). These cells exhibit reduced polyfunctionality, particularly in IL-2 and IFN- $\gamma$ production, compared to Gzma-FM- cells.
Evidence for Bifurcation: At early memory stages (days 29–35), the majority of central memory T cells ( $T_{CM}$ ) are Gzma-FM-, supporting the bifurcation model where memory precursors arise from cells that never expressed Gzma.
Evidence for Linear Dedifferentiation: Over time, the proportion of Gzma-FM+ cells within the $T_{CM}$ pool steadily increases. By day 130 post-infection, approximately 30% of $T_{CM}$ cells in the spleen and lymph nodes are Gzma-FM+. These "dedifferentiated" cells regain Tcf-1 expression, CD62L expression, and the capacity to produce IL-2, reaching functional parity with Gzma-FM- memory cells.
Recall Competence: Gzma-FM+ memory cells isolated at late time points (day 180) exhibit robust recall expansion and distribution comparable to Gzma-FM- counterparts upon secondary challenge. However, they show a tendency to differentiate more rapidly into KLRG1+ and CX3CR1+ effector phenotypes during the secondary response, suggesting epigenetic imprinting from their primary effector history.

Significance and Claims
The paper claims to resolve the dichotomy between linear and bifurcation models by demonstrating that both models are operational following acute viral infection, but at different temporal scales:

Early Phase: The bifurcation model dominates. Early memory precursors arise from cells that never expressed Gzma, establishing the initial memory pool.
Late Phase: The linear model progressively contributes. Terminal effector T cells (Gzma+), previously thought to be a fixed lineage, can dedifferentiate into functional, long-lived memory T cells over time.

The authors conclude that the long-term memory pool is a heterogeneous mixture of cells with distinct differentiation histories. The Gzma-FM system reveals that Gzma expression marks a state of terminal cytotoxicity that is reversible under specific conditions, allowing these cells to eventually join the memory compartment. This finding suggests that the immune system utilizes a dual strategy: rapid establishment of memory via bifurcation and a slower, progressive contribution from terminal effectors to ensure balanced and durable immune protection.

Limitations and Modesty
The authors note that the dedifferentiation of Gzma+ cells is a slow and progressive process, explaining why it was missed in earlier "snapshot" studies. They also acknowledge that while Gzma-FM+ memory cells are functionally competent, they retain a "memory" of their effector state, leading to a distinct secondary effector phenotype (skewed toward terminal effectors) compared to Gzma-FM- derived cells. The study does not claim that Gzma+ cells are the primary source of memory, but rather a significant, delayed contributor that complements the early bifurcated population.