Exploiting HLA-II Promiscuity via Peptide Terminal Overhang Recognition for Pan-Allelic and Tumor-Selective AML Immunotherapy

This study develops a pan-allelic bispecific T cell engager (BiTE) that targets an MPO-derived peptide's N-terminal overhang presented promiscuously by HLA class II molecules, enabling potent and selective killing of acute myeloid leukemia cells while sparing normal myeloid cells.

The Gist

The Big Picture: A New Way to Hunt Cancer

Imagine the human body is a massive city, and the immune system is the police force. Usually, the police can only catch criminals (cancer cells) if the criminals are wearing a specific, unique "wanted poster" on their chest that the police recognize.

However, cancer cells are tricky. They often hide their "wanted posters" inside their buildings, making them invisible to the police. Furthermore, every person in the city has a slightly different police badge (called an HLA molecule). A police officer trained to recognize a poster on Badge Type A might not recognize the same poster on Badge Type B. This means current cancer treatments often only work for a small group of people who happen to have the right badge.

This paper describes a breakthrough strategy to catch cancer cells that works for almost everyone, regardless of their badge type, and does so without accidentally arresting innocent people.

---

The Problem: The "Locked Room" vs. The "Open Window"

The Old Way (HLA Class I):
Think of the standard way the immune system sees cancer as a "locked room." The cancer cell puts a tiny piece of its internal protein (a peptide) into a small, closed window (the HLA Class I molecule).

• The Catch: The window is very small and shaped specifically for one person's badge. If the badge doesn't match perfectly, the police can't see the piece of protein inside. It's like trying to fit a square peg into a round hole.

The New Discovery (HLA Class II):
The researchers looked at a different kind of window on the cancer cell called HLA Class II.

• The Difference: This window is like a long, open hallway. It can hold much longer pieces of protein. Because the hallway is open at both ends, the protein sticks out on both sides, like a long scarf hanging out of a coat pocket.
• The Opportunity: Even though the middle of the scarf might be tucked into different parts of the hallway depending on the person's badge, the ends of the scarf (the overhangs) are always hanging out in the open. These ends look the same regardless of the badge type.

The Solution: The "Scarf-Grabbing" Antibody

The researchers focused on a specific protein found in Acute Myeloid Leukemia (AML), a type of blood cancer, called MPO.

1. Finding the Target: They found a long piece of the MPO protein (called MPO100-132) that sticks out of the "hallway" window on cancer cells.
2. The Magic Key: They created a special antibody (a smart weapon) named 146D5.
   • Instead of trying to grab the whole scarf and the badge together (which is hard because every badge is different), this antibody was designed to only grab the very tip of the N-terminal overhang (the end of the scarf).
   • Because the tip of the scarf looks the same whether it's hanging out of a DP4 badge, a DR1 badge, or a DR15 badge, this one antibody works on many different types of people. It is "pan-allelic" (works for all).

The Safety Feature: Why It Doesn't Hurt Healthy People

You might ask: "If MPO is in all my blood cells, won't this antibody kill my healthy white blood cells too?"

This is the cleverest part of the study.

• The "Scarf" is a Construction Zone: The specific piece of the MPO protein that the antibody grabs (the N-terminal overhang) is actually part of the construction scaffolding used to build the protein.
• In Healthy Cells: When healthy blood cells finish building the MPO protein, they throw away the scaffolding. The "scarf tip" is gone. The antibody has nothing to grab onto.
• In Cancer Cells: AML cancer cells are like rushed construction sites. They are so busy and chaotic that they often leave the scaffolding (the pro-peptide) attached to the finished building. The "scarf tip" is still hanging there.
• The Result: The antibody sees the cancer cell, grabs the scaffolding tip, and calls in the T-cells (the police) to destroy it. It ignores the healthy cells because they have already cleaned up their construction site.

The Weapon: The "BiTE" (Bispecific T-cell Engager)

To make this antibody a weapon, the researchers turned it into a BiTE.

• Imagine a double-sided tape.
• Side A: The 146D5 antibody that grabs the cancer cell's "scarf tip."
• Side B: A hook that grabs a T-cell (the police officer).
• The Action: When you inject this into a patient, it physically links the T-cell to the cancer cell. Once they are taped together, the T-cell immediately attacks and kills the cancer.

The Results: A Victory in the Lab

The researchers tested this in the lab and in mice:

1. Broad Reach: It successfully killed leukemia cells from many different patients, even those with different HLA badges.
2. Precision: It ignored healthy blood cells, proving it is safe.
3. Power: In mice with leukemia, this treatment stopped the cancer from growing and helped the mice live much longer.

Summary Analogy

Imagine a city where criminals (cancer) wear a long, colorful scarf (MPO protein).

• Old Police: Could only see the scarf if it was tied in a very specific knot that matched their specific badge. If the knot was different, they missed the criminal.
• New Strategy: The researchers realized that no matter how the scarf is tied, the fringe at the end always hangs loose.
• The New Weapon: They built a magnet (the antibody) that only sticks to the fringe.
• The Safety: Only the criminals are wearing the scarf with the fringe still attached. The good citizens (healthy cells) have already cut off the fringe.
• The Outcome: The magnet grabs the criminal's fringe, pulls a police officer (T-cell) right next to them, and the criminal is arrested. It works on every criminal in the city, regardless of their badge, and never touches the innocent citizens.

This paper suggests a new, powerful way to treat leukemia that could work for a much wider range of patients than current therapies.

Technical Summary

1. Problem Statement

Current antibody-based immunotherapies targeting peptide-MHC (pMHC) complexes, particularly TCR-mimicking (TCRm) antibodies, face a significant limitation: HLA restriction. Most TCRm antibodies recognize peptides presented by a single specific HLA allele (e.g., HLA-A*02:01), limiting their applicability to a small subset of the patient population.

• The Challenge: While HLA Class I (HLA-I) presents short, fixed-length peptides in a closed groove, HLA Class II (HLA-II) presents longer peptides (13–25 amino acids) with open-ended binding clefts. This allows for promiscuous presentation, where a single peptide can bind multiple HLA-II alleles via different binding registers.
• The Gap: Despite the potential of HLA-II promiscuity, it remains underexploited for immunotherapy. Furthermore, targeting myeloid antigens like Myeloperoxidase (MPO) in Acute Myeloid Leukemia (AML) carries a risk of on-target, off-tumor toxicity against normal myeloid cells (neutrophils/monocytes) which also express MPO.

2. Methodology

The authors employed a multi-step strategy to identify a promiscuous epitope, generate a cross-reactive antibody, and validate its therapeutic potential:

• Epitope Discovery:
  • Analyzed transcriptomic data from AML cohorts to confirm co-expression of MPO and HLA-II.
  • Used in silico prediction (NetMHCII) to identify MPO-derived peptides capable of binding multiple HLA-II alleles (focusing on DP4, DR, and DQ).
  • Validated endogenous processing and presentation using peptide-loaded HLA-II dimers and TCR cloning from AML patients.
• Antibody Generation & Characterization:
  • Immunized mice with N-terminal (MPO100-112) and C-terminal (MPO122-132) overhang peptides derived from the promiscuous MPO100-132 sequence.
  • Screened hybridomas to identify clones (specifically 146D5) that recognize the peptide overhang independent of the HLA scaffold.
  • Used Alanine/Serine scanning, competitive binding assays, and AlphaFold structural modeling to map the epitope and confirm it resides on the exposed N-terminal overhang, distinct from the HLA-binding core.
• Therapeutic Engineering:
  • Engineered the 146D5 scFv into a Bispecific T-cell Engager (BiTE) fused with an anti-CD3 scFv (OKT3).
• Functional Validation:
  • Tested binding and cytotoxicity against K562 cells engineered with various HLA-II alleles and MPO.
  • Evaluated activity against primary AML samples and normal human myeloid cells (neutrophils/monocytes).
  • Assessed cross-reactivity with homologous peptides (e.g., Eosinophil Peroxidase, EPX).
• In Vivo Efficacy:
  • Utilized NSG mouse xenograft models (Kasumi-1 and Nomo-1) treated with 146D5 BiTE and adoptively transferred polyclonal T cells.

3. Key Contributions

• Novel Targeting Strategy: The paper introduces a paradigm shift from HLA-restricted to allele-agnostic targeting by exploiting the N-terminal overhang of HLA-II presented peptides. This overhang remains structurally accessible and conserved across different HLA-II alleles, even when the peptide core binding register shifts.
• Pan-Allelic Recognition: Demonstrated that a single antibody (146D5) can recognize the same MPO-derived peptide presented by diverse HLA-II alleles (including DP4, DR1, DR15, and DQ6.1) with high affinity (~25 nM).
• Tumor Selectivity Mechanism: Identified a unique safety window where the therapeutic target (MPO100-132) is derived from the MPO propeptide. While AML cells accumulate immature MPO precursors (apoproMPO/proMPO) due to processing defects, normal myeloid cells efficiently process these precursors into mature MPO, removing the propeptide. Consequently, the target epitope is absent on normal cells despite high MPO expression.

4. Key Results

• Promiscuous Binding: The MPO100-132 peptide binds promiscuously to multiple HLA-II alleles. Structural modeling confirmed that while anchor residues differ between alleles (e.g., DP4 vs. DR15), the N-terminal overhang (residues Y101, F102, P105, V106, A107) remains exposed.
• Antibody Specificity: The 146D5 antibody specifically binds the N-terminal overhang. Mutations in the overhang residues abolished binding, while mutations in the HLA-binding core did not affect antibody recognition (though they affected peptide loading).
• BiTE Efficacy: The 146D5 BiTE induced potent, concentration-dependent T-cell cytotoxicity against:
  • MPO+ AML cell lines (ME-1, Kasumi-1, Nomo-1).
  • Primary AML patient samples with diverse HLA-II backgrounds.
  • Cells expressing various HLA-II alleles (DP, DR, DQ).
• Safety Profile:
  • Normal Myeloid Cells: 146D5 BiTE showed no cytotoxicity against normal neutrophils or monocytes, even when these cells expressed HLA-II and MPO.
  • Mechanism of Safety: Western blotting confirmed the presence of immature apoproMPO/proMPO in AML cells but its absence in normal cells.
  • Cross-reactivity: While the antibody recognized a homologous peptide from Eosinophil Peroxidase (EPX), EPX is expressed in eosinophils which lack HLA-II, preventing off-target killing.
• In Vivo Success: In xenograft models, 146D5 BiTE treatment significantly suppressed tumor growth and prolonged survival compared to controls, demonstrating potent anti-leukemic activity.

5. Significance

• Overcoming HLA Restriction: This study provides a robust blueprint for developing "pan-allelic" immunotherapies. By targeting peptide overhangs rather than the peptide-HLA interface, therapies can potentially treat a much broader patient population without requiring HLA matching.
• Safety in Myeloid Targets: It solves a major hurdle in targeting myeloid antigens (like MPO) by leveraging the differential processing of immature protein precursors in cancer vs. normal cells. This offers a high therapeutic index for AML.
• Broader Applicability: The strategy of targeting promiscuously presented HLA-II overhangs is not limited to AML or MPO. It suggests a generalizable approach for targeting other intracellular antigens in solid tumors and hematological malignancies that express HLA-II, potentially expanding the "druggable" proteome for antibody-based therapies.
• Clinical Potential: The development of a BiTE that is both broadly applicable (pan-allelic) and highly selective (sparing normal tissue) represents a significant step forward in the next generation of T-cell retargeting immunotherapies.