The Week in Research: mRNA Oncology, Ambient Direct Air Capture, and 'Flying Qubits'

This week's breakthroughs span the boundary between molecular medicine, climate technology, and quantum hardware. From reprogramming the immune system to fight cancer without leaving the patient's body, to a room-temperature carbon capture material that runs on waste heat, to topologically protected quantum information highways that could make fault-tolerant quantum computers achievable — the week of July 24, 2026 delivered three exceptional research advances across three of science's most consequential frontiers.

---

🔬 mRNA Technology Expands Into In Vivo Oncology

The Problem with Existing CAR-T Cell Therapy

Traditional CAR-T (chimeric antigen receptor T-cell) therapies have achieved remarkable results against blood cancers — but they face two structural problems that have limited their impact:

The existing CAR-T process:

The problem: solid tumours (lung, pancreatic, ovarian, colorectal) present an additional challenge — the tumour microenvironment actively suppresses T-cell function, so even successfully engrafted CAR-T cells are "exhausted" before reaching the tumour.

What the Dana-Farber / BioNTech Study Found

The approach — in vivo mRNA reprogramming:

Instead of extracting and modifying T-cells, the researchers developed targeted lipid nanoparticles (LNPs) carrying mRNA that reprograms T-cells directly inside the body:

Key technical innovations:

1. Selective T-cell targeting: The LNPs are coated with a CD3ε-binding antibody fragment — this makes them specifically dock onto CD3ε, a protein present on all T-cells but not on other cells. Only T-cells internalise the LNPs.

2. Temporary CAR expression: The mRNA encodes a CAR receptor that remains active for 5–10 days (the lifespan of a single mRNA strand before cellular ribosomes degrade it). This avoids permanent genetic modification (reducing off-target cancer risk) and allows re-dosing with modified CAR sequences as tumour antigens mutate.

3. Tumour antigen targeting: In the preclinical models, the study used mRNA encoding a CAR targeting HER2 (overexpressed in breast, lung, and gastric cancers) and EGFRvIII (a brain tumour-specific antigen). Both showed complete tumour elimination in 78% of treated mice at Day 30.

4. Co-delivery of immune stimulants: Each LNP co-delivers mRNA encoding IL-12 (an immune activator) alongside the CAR sequence. IL-12 counteracts the suppressive tumour microenvironment — specifically blocking the PD-1/PD-L1 checkpoint that tumours use to silence T-cells.

Why this changes oncology economics:

• Eliminates the ex vivo manufacturing bottleneck: no leukapheresis, no cleanroom cell processing, no 4-6 week wait
• LNPs are already manufactured at industrial scale for mRNA COVID vaccines — the same production infrastructure applies
• Estimated cost (BioNTech's internal projection for eventual commercial scale): $15,000–$30,000 per dose vs $400,000–$700,000 for ex vivo CAR-T
• Treatment timeline: injection to active immune response in 48–72 hours vs 5–8 weeks

Current stage: Preclinical (mouse models). BioNTech has filed an IND (Investigational New Drug) application with the FDA for a Phase 1 dose-escalation trial targeting HER2+ solid tumours. Expected trial start: Q1 2027.

---

🌱 Ambient-Temperature Direct Air Capture — The MOF Breakthrough

Why DAC Energy Cost Has Been the Blocking Problem

Current commercial DAC plants (Climeworks' Mammoth plant in Iceland; Carbon Engineering's plant in Texas) use amine-based sorbents that require heating to 80–120°C to release (desorb) captured CO₂. This energy requirement is the dominant cost:

Current cost: approximately $400–$1,000 per tonne of CO₂ removed. The Biden-era (and continued Trump-era) DAC target is $100/tonne to make it economically viable alongside other climate interventions.

The Berkeley/PSI Metal-Organic Framework

The material: The new MOF (designated PSI-MOF-74-MNO3 in the paper, a variant of the Mg₂(dobdc) family) achieves three simultaneous advances:

Why the low desorption temperature matters: At 42–48°C, the MOF can release its CO₂ using:

• Industrial waste heat (most industrial processes produce waste heat at 40–80°C that is currently vented to the atmosphere)
• Concentrated solar thermal (simple flat-plate solar collectors produce 50–70°C water)
• Data centre cooling water (hyperscale data centres exhaust coolant at 40–60°C)

This means DAC could be co-located with industrial facilities or data centres and run on zero-cost waste heat — fundamentally changing its economics.

Modelling by Berkeley's Kammen Group: Assuming the MOF reaches commercial production at $20/kg (comparable to industrial adsorbents), pairing it with free waste heat from data centres, the modelled capture cost falls to $68–$95/tonne — below the $100/tonne threshold.

---

💻 "Flying Qubits" — Topological Quantum Interconnects

The Decoherence Problem in Quantum Computing

Quantum computers must maintain qubits in superposition (simultaneously 0 and 1) to perform calculations. Environmental noise — thermal vibrations, electromagnetic interference, even cosmic rays — "decoheres" qubits, collapsing them to classical states and destroying the calculation. This is the primary barrier to scaling quantum computers beyond ~1,000 physical qubits.

Coherence times of current qubit technologies:

The Cologne University Experiment

What chiral Majorana edge states are: Majorana fermions are particles that are their own antiparticles — meaning they cannot be "annihilated" by environmental interactions in the same way ordinary particles can. Crucially, their quantum state is stored non-locally across the edge of the material, not at a single point — so local perturbations (noise) cannot corrupt the information.

What the experiment achieved: The team sandwiched a magnetically doped topological insulator (Cr-doped (Bi,Sb)₂Te₃) between a superconductor (Nb). Under a specific magnetic field, they observed:

• One-dimensional conduction along the material edge (confirmed via resistance measurements)
• Quantized conductance of e²/2h (the exact value predicted for a single chiral Majorana mode)
• The edge current flows in only one direction — it cannot backscatter (a noise source that limits conventional quantum signals)

The "flying qubit" concept: A chiral Majorana edge mode acts like a protected one-way highway for quantum information — a qubit encoded in this mode flies along the edge without decoherence. Two nodes of a quantum processor could exchange quantum information via these modes as interconnects, solving the wiring problem in large quantum chips.

What's still needed:

• Demonstrate braiding (using two Majorana modes to perform a topological quantum gate) — this has not yet been achieved; the Cologne experiment shows the edge modes exist but doesn't yet braid them
• Operate at temperatures above ~100 mK (current requirement) — still requires dilution refrigerators
• Scale to useful gate counts

Timeline: Most quantum computing experts estimate topological qubits will be competitive with superconducting qubits by 2030–2032, with Microsoft's "Station Q" research group working in parallel on a similar approach (using InAs nanowires rather than topological insulators).

---

📌 The Bottom Line

• mrna-oncology-in-vivo: Dana-Farber + BioNTech (Nature): CD3ε-targeting LNPs deliver temporary CAR mRNA + IL-12 directly into T-cells in vivo; HER2 + EGFRvIII CARs = 78% complete tumour elimination in mice (Day 30); eliminates leukapheresis + ex vivo manufacturing (5-8 weeks → 48-72hrs); estimated cost $15K-$30K/dose vs $400K-$700K traditional CAR-T; BioNTech Phase 1 IND filed → Q1 2027 trial start.
• ambient-dac-mof: Berkeley + Paul Scherrer Institute (Science): PSI-MOF-74-MNO3 — CO₂ capture 3.8 mmol/g (vs 2.1), desorption at 42-48°C (vs 80-120°C), 320-480 kWh/tonne (vs 1,500-2,000), 96% stability over 1,000 cycles; desorption sources: industrial waste heat + solar thermal + data centre cooling water (all free); modelled cost at scale: $68-$95/tonne (below $100/tonne viability threshold).
• flying-qubits-majorana: U of Cologne (Nature Physics): Cr-(Bi,Sb)₂Te₃/Nb junction → quantized conductance e²/2h (single chiral Majorana mode confirmed); one-directional backscatter-immune edge current = topologically protected quantum information channel; Majorana coherence = theoretically infinite (non-local state storage); next step: braiding for topological gates (not yet achieved); competitive timeline vs superconducting qubits: 2030-2032 (Microsoft Station Q parallel programme).

---

📬 Stay Updated

Get the week's most important research breakthroughs delivered to your inbox. Subscribe to our free newsletter →

---

Disclosure: This post contains affiliate links. If you purchase through our links, we earn a small commission at no extra cost to you. We only recommend products we believe in.