Here's the assumption baked into most obesity-inflammation research: it's the chronic fat accumulation that wrecks your immune system. Gain weight, develop metabolic dysfunction, watch your inflammatory disease spiral. The diet is just the vehicle. The obesity is the problem.
This paper torches that assumption.
Lean mice. No weight gain. Just four days of high-fat chow timed to coincide with the start of a skin flare. That brief window is enough to permanently rewire the immune response, creating a memory that makes every future flare worse, even after the mice go back to eating normally for weeks.
The implications reach well beyond psoriasis. If transient dietary choices during an inflammatory episode can reprogram long-term disease trajectory, then the question of what you eat during a flare matters in a way nobody has been accounting for.
The experimental setup is clean. Mice on a low-fat diet (LFD, 10 kcal% fat) serve as controls. One group gets one week of high-fat diet (HFD, 60 kcal% fat) before psoriasis induction via topical imiquimod (IMQ). Another group gets ten weeks of HFD, long enough to become obese. Then everyone gets the same inflammatory challenge.
Disease severity is tracked daily by measuring ear thickness with a digital micrometer. Immune populations in the lesional skin are quantified by multi-color flow cytometry, gating on CD45+ CD3ε+ CD4+ FOXP3- conventional T cells and their cytokine output (IL-17A, IL-17F, IL-22). The gating strategy is shown in Extended Data Fig. 1, and it's thorough enough that you can trust the population calls.
The one-week HFD group never gains meaningful weight. These are lean mice by any clinical definition. And yet, at peak disease, their ear swelling is 2.1-fold above LFD controls. The obese mice on ten weeks of HFD? 2.3-fold. Statistically, those two numbers are not different from each other.
The skin histology matches: both HFD groups show substantial expansion of the dermis and epidermis. TH17 cells, the pathogenic drivers of psoriatic inflammation, are elevated in both groups. Neutrophils pile up. Regulatory T cells drop. The immune profile of a lean mouse on one week of HFD is nearly indistinguishable from an obese mouse.
result
Four days of HFD, timed to overlap with the start of the immune response, is sufficient to drive the full inflammatory phenotype. Feed the mice HFD outside that window, before or after the flare begins, and disease severity stays at LFD levels. The diet only matters when it coincides with the immune system waking up.
So what cell type is translating the dietary signal into immune damage? The authors ran a systematic knockout screen. Rag1-/- mice (no T or B cells) are completely resistant to HFD-induced exacerbation. That points to adaptive immunity. Knocking out gamma-delta T cells doesn't help. Knocking out CD8+ T cells doesn't help. Only removing CD4+ T cells abolishes the HFD effect entirely.
The proof-of-concept experiment is elegant: take bulk CD4+ T cells from wild-type spleens, transfer them into Rag1-/- mice, let them reconstitute for eight weeks, then run the HFD-plus-IMQ protocol. The reconstituted mice show the full HFD-driven disease exacerbation. CD4+ T cells alone are sufficient to carry the dietary signal into inflammatory damage.
CD4+ T cells are the essential link between a high-fat diet and worsened inflammatory disease. No CD4+ T cells, no diet effect, regardless of how much fat the mouse eats.
result
Single-cell RNA sequencing on CD4+ T cells sorted from lesional skin identifies eight clusters: TH1, TH17, proliferating TH17, and four Treg populations. When you compare HFD versus LFD cells within the TH17 clusters, one gene rises to the top of the differential expression analysis: Il1r1, the gene encoding the IL-1 receptor, is the most significantly upregulated gene in TH17 cells from HFD-fed mice.
Flow cytometry confirms this at the protein level. IL1R1 mean fluorescence intensity on RORγt+ TH17 cells is roughly 1.6-fold higher in HFD mice (3,779 vs 2,313 MFI). The cells expressing IL1R1 also score higher on published TH17 pathogenicity gene signatures. This is not a bystander receptor. It's marking the most dangerous cells in the lesion.
The functional test is decisive. Mice with T-cell-specific deletion of IL1R1 (Cd4Cre Il1r1fl/fl) show no difference in disease severity between HFD and LFD. The HFD effect is completely gone. The receptor on the T cell is required.
Upstream of the T cells, bulk RNA-seq of lesional ear skin shows a ~3.9-fold increase in Il1b expression in HFD mice, accompanied by upregulation of Nlrp3. The NLRP3 inflammasome processes pro-IL-1β into active IL-1β, which then signals through IL1R1 on TH17 cells. Western blots confirm elevated active caspase-1 (p20) in the lesional skin-draining lymph nodes of HFD mice, but not in the contralateral, unchallenged side. The activation is local, not systemic.
Nlrp3-/- mice on HFD show no disease exacerbation. Topical application of MCC950, a well-characterized small-molecule NLRP3 inhibitor, brings HFD-driven ear swelling back to LFD levels and reduces TH17 cells and neutrophils in the skin. The pathway is NLRP3 in innate cells, IL-1β release, IL1R1 signaling on TH17 cells, amplified inflammation.
The NLRP3-IL1R1 axis is the molecular bridge between a high-fat diet and pathogenic TH17 cell expansion. Block NLRP3 topically during the flare, and the diet effect disappears.
result
The most striking experiment in the paper is the recurrence model. Mice get one week of HFD plus six days of IMQ (the initial flare). Then the diet switches back to LFD. After roughly 25 days of recovery, during which ears look completely normal by gross examination, histology, and thickness measurement, the mice get a second IMQ challenge.
The mice that had HFD during their first flare (HFDDODO) show a roughly 2-fold greater ear thickness increase during the second flare compared to mice that were on LFD throughout. TH17 cell infiltration and neutrophil accumulation are both higher. The Tconv cells in the rechallenged ears still express elevated IL1R1. The immune system remembered the diet, even though the diet is long gone.
Two controls sharpen the interpretation. First, mice given HFD for one week without any IMQ challenge, then returned to LFD for two weeks before IMQ, show no difference from LFD controls. The diet alone, without an inflammatory trigger, does not create memory. Second, the memory is tissue-specific: only the previously challenged ear shows exacerbated recurrence. The unchallenged contralateral ear behaves normally. Blocking lymphocyte egress with FTY720 during the second flare does not prevent the exacerbation, pointing to tissue-resident cells rather than circulating ones.
Recovered ears from HFDDODO mice contain approximately 2.3-fold more resident memory CD4+ T cells (Trm, defined as CD4+ FOXP3- CD62Llow CCR7low CD44high CD69+) than LFD controls. Anti-CD4 antibody treatment during the second flare abolishes the HFDDODO effect. The memory lives in the tissue, it's carried by CD4+ T cells, and it was written during the first flare.
The therapeutic window finding is the practical punchline: topical MCC950 applied during the first flare prevents the exacerbated recurrence weeks later. The same treatment applied only during the second flare provides no comparable protection. You have to intervene at onset, or the memory is already set.
model
The central reframe here is worth sitting with. Obesity research has treated the chronic metabolic state as the primary immune risk factor. This paper argues that the acute dietary signal during an inflammatory event is doing independent, durable damage, and that lean individuals are not protected from it.
From a bench perspective, the NLRP3-IL1R1 axis is already a druggable target. MCC950 is a well-characterized tool compound, and topical delivery to inflamed skin is mechanistically plausible. The finding that the therapeutic window closes after the initial flare is a real constraint, but it also defines a clear clinical question: can NLRP3 inhibition during a first psoriatic episode reduce the severity of future flares?
One honest caveat: all experiments used male mice, and the IMQ model, while well-validated, is an acute induction system rather than a spontaneous chronic disease. Whether the DODO effect translates to human psoriasis, or to other inflammatory conditions where diet is a known but poorly understood trigger, remains an open question. The mechanistic pathway is compelling enough to take seriously, but the leap from mouse ear to human skin requires its own evidence.
What this paper does establish, cleanly and with good genetic controls, is that the timing of dietary exposure relative to immune activation is a variable that matters, and that ignoring it because a patient isn't obese is probably a mistake.