
(Prostock-studio/Shutterstock)
Scientists Question Whether the Immune System Really Starts Allergies
In A Nutshell
- Researchers propose that allergies may begin when barrier cells lose the ability to protect their own mitochondria, with the immune system reacting afterward.
- Gene data showed the clearest signs in eosinophilic esophagitis, where mitochondria-clearing genes were lower in all 13 patient groups.
- Missing in established asthma and nasal polyps, the pattern was only partial in hay fever, and psoriasis and COPD showed similar signs.
- This remains a hypothesis, and the authors say infant studies before allergies appear are the key next step.
Allergies may not start in the immune system, after all. A new theory published in the journal Allergy argues that the hidden cause may be a power failure inside the cells that form the body’s protective barriers, with the immune system joining in afterward. The clue is relapse: stop an effective immune-blocking drug for conditions like asthma or chronic sinus disease with nasal polyps, and symptoms often roar back within months. The suspect is mitochondria, the tiny structures inside cells that produce energy.
A team of 10 international researchers proposes that this reversal may apply to eczema, asthma, hay fever, and some food allergies. In their model, cells lining the skin, esophagus, nose, and airways lose the ability to protect their own mitochondria, weakening the barrier, and the immune system gets dragged in later. That order could explain why many treatments control symptoms without delivering a lasting cure.
Allergic disease has risen too fast for genes alone to explain, and food allergy now affects roughly one child in 20 in industrialized countries. For 60 years, the standard account has run one way: immune cells mistake harmless things like peanuts for threats, and the inflammation then damages the skin or gut lining. The authors cite two more clues pointing the other direction. A sizable minority of patients never fully respond even to drugs that block a key immune pathway. And babies with leakier skin at 3 months had more than three and a half times the odds of developing early signs of allergy by 6 months, according to an earlier study the paper cites.
Gene Data Tie Mitochondria and Allergy Problems to the Esophagus and Skin
Rather than recruiting patients, the team mined existing data: 116 patient-versus-control comparisons from earlier studies, covering 3,776 patients and 2,164 controls, including some with the skin disease psoriasis and the lung disease COPD as non-allergic comparisons. Tissue came from skin, esophagus, nose, airways, and blood.
Of 212 comparisons, only nine survived a standard safeguard against false positives. Two merely confirmed the immune reaction was present. Five involved how esophageal cells handle energy, and two involved an internal alarm system that reacts to stray genetic material, such as mitochondrial DNA escaping into a cell’s interior. That escape was not measured directly.
Esophageal tissue from people with eosinophilic esophagitis, a chronic allergic condition of the esophagus, gave the clearest results. In all 13 patient groups, genes that clear out damaged mitochondria ran at lower levels. The cells did not seem short on basic energy output, but they lost key defenses that protect their mitochondria. Two genes in that alarm system also behaved differently in allergic tissue. Comparing them ranked the esophagus, eczema skin, and nasal lining in hay fever above psoriasis, asthma, and COPD. Only the esophagus result survived the false-positive safeguard, and the comparison covered tissue types, not individual patients.
In one set of samples from eczema patients’ skin that looked normal and showed no visible inflammation, genes that build the skin barrier and antioxidant defenses were lower than in healthy skin, while a gene that blocks those defenses was higher. Pooling nine sample sets mostly pointed the same way without reaching statistical significance, so the fault may exist before flare-ups, but the data cannot prove it.
Asthma and Nasal Polyps Lack the Same Mitochondria Pattern
Not every condition fit. The pattern was missing in nasal polyps and established adult asthma, and appeared only under certain conditions in hay fever. Food allergy was the weakest case, since its datasets came only from blood. The same mitochondrial patterns also appeared in psoriasis and COPD, so they may reflect chronic tissue damage in general rather than allergy specifically.
Most Genetic Scans Skip Mitochondria, Which May Hide the Allergy Link
Researchers say blind spots in study design explain why decades of genetic research missed mitochondria. Standard DNA-testing chips barely cover mitochondrial DNA, and no large genetic study in three decades has linked it to allergy. Many of the proposed cell events also happen after proteins are made, through chemical changes that leave no DNA mutation for a genetic scan to find. In the authors’ words, “The targets were not absent. They were unmeasured.”
Allergy Drugs Mostly Calm Immune Cells and Skip the Proposed Energy Fault
Almost every approved allergy therapy works downstream, calming the immune response once it is underway. Relapse can also stem from leftover allergen exposure or lingering immune memory, and no trial of an immune-blocking drug has checked whether the proposed energy fault remains afterward.
Study lead Peter K. Smith, a clinical immunologist at Griffith University, said: “We have spent our careers managing the smoke and for the first time we think we can see the fire, and it is not where everyone was looking.” He added that the next step is “to measure these cells directly in infants before the allergy develops.”
This paper does not prove that failing mitochondria are the reason, but it earns them a direct look before the immune system takes all the blame.
Disclaimer: This article summarizes a published research paper for general information and is not medical advice. The theory described is a hypothesis that has not been proven or tested as a treatment. Anyone with questions about allergies or their care should talk with a qualified health care professional.
Paper Notes
Limitations
This is a review proposing a hypothesis. It does not test whether restoring mitochondrial function would prevent or cure allergies. Most patient samples came from tissue that was already inflamed, so the data cannot show whether the cellular energy problems came before disease or resulted from it. The authors note that even normal-looking skin has been exposed to years of environmental influences and treatment. Two central ideas in the model, damage to a fat molecule in mitochondrial membranes and mitochondrial DNA leaking into the cell’s interior, were inferred rather than measured, and gene activity is an indirect stand-in for actual mitochondrial function. The comparison of two alarm-system genes (STING1 and MAVS) was built after STING1 had already stood out, and only its esophagus value survived the false-positive safeguard. The patient total includes people with psoriasis and COPD, who served as non-allergic comparison groups. Food allergy data came only from blood, and the nasal polyp analysis included just three patient groups. Broad suppression of mitochondrial activity across all allergic tissues was not supported, and similar patterns appeared in psoriasis and COPD. The authors state that prospective and experimental studies must establish timing and cause.
Funding and Disclosures
Funding is listed as none: the authors state they have nothing to report. Open access publishing was facilitated by Griffith University through its agreement with Wiley and the Council of Australasian University Librarians. The analysis used only publicly deposited datasets, so no ethics approval was required. According to the paper, the data were analyzed with a purpose-built large-language-model workflow implemented in Anthropic Python code, while the manuscript itself was author-written, and the gene list and scoring rules were set before any dataset was seen. Several authors reported industry ties. Peter K. Smith disclosed investigator-initiated funding from GSK and Sanofi, speaker fees from Danone, Nestlé, and Medscape, advisory board service for Viatris, and a provisional patent application relating to diagnosis and treatment based on the proposed cellular fault. David Price disclosed advisory, consulting, grant, and speaking relationships with numerous drug makers, including AstraZeneca, GlaxoSmithKline, Novartis, and Sanofi Genzyme, plus majority ownership stakes in Optimum Patient Care Ltd. and the Observational and Pragmatic Research Institute Pte Ltd. Other co-authors reported grants, consulting fees, or honoraria from companies such as GSK, Sanofi, Regeneron, Novartis, Danone, and Nestlé Health Sciences. Cezmi A. Akdis is editor-in-chief of Allergy. Stanley Du Preez declared no conflicts. Full disclosures appear in the paper.
Publication Details
Title: The Origins of Allergy: An Outside-In, Bioenergetic Hypothesis
Authors: Peter K. Smith (Griffith University, Gold Coast, Australia), Nicholas P. West (Griffith University), Stanley Du Preez (Griffith University), Hugh A. Sampson (Icahn School of Medicine at Mount Sinai, New York), Peter W. Hellings (University Hospitals Leuven, Belgium), David Price (Observational and Pragmatic Research Institute, Singapore, and University of Aberdeen, UK), Liam O’Mahony (University College Cork, Ireland), Carina Venter (Wake Forest University, North Carolina), Cezmi A. Akdis (Swiss Institute of Allergy and Asthma Research, University of Zurich), Gil Yosipovitch (University of Miami Miller School of Medicine) Journal: Allergy, 2026 (published by the European Academy of Allergy and Clinical Immunology and John Wiley & Sons Ltd.)
Article type: Review, open access Dates: Received July 10, 2026; revised September 11, 2026; accepted September 24, 2026 DOI: https://doi.org/10.1111/all.70532 Corresponding author: Peter K. Smith ([email protected])







