Immune Alarms Years Before Pain: Blood Signatures in At-Risk People Reveal How Rheumatoid Arthritis Begins

Imagine a smoke alarm that can hear the faintest hiss of a smoldering wire long before anyone smells smoke. That is the promise behind searching for biological signals that precede rheumatoid arthritis, an autoimmune disease that often goes diagnosable only after joint damage begins. If clinicians could reliably detect the disease process earlier, targeted interventions might delay or prevent the pain and disability that follow. A new prospective, multiomic study led by He and colleagues tracks people who are positive for anticitrullinated protein antibodies, or ACPAs, but who do not yet have swollen joints. The team finds clear signs of inflammation and a surprising reprogramming of T and B lymphocytes long before clinical disease appears, offering a roadmap for earlier detection and smarter prevention.

The study at a glance: what the researchers did and who they watched

Researchers enrolled 45 clinically healthy, ACPA-positive individuals deemed “at risk” (ARI), 11 patients with early clinical rheumatoid arthritis (ERA), and multiple ACPA-negative healthy control groups. The at-risk group was followed prospectively with repeated blood sampling and deep immune profiling that combined plasma proteomics, single-cell RNA sequencing of peripheral blood mononuclear cells, immune repertoire analyses, chromatin accessibility assays, and functional cell stimulation assays. Over the study period, 16 of the 45 at-risk participants, about 36 percent, progressed to clinically evident RA, allowing a rare window into the immune events that accompany conversion.

The team pooled these data into integrated, interactive resources so other scientists can explore specific cell types, gene programs, and protein signatures discovered in the study.

What the researchers found

Systemic inflammation is already present before the first swollen joint

Even when individuals had no clinical signs of synovitis, their plasma showed a broad inflammatory signature. The scientists identified hundreds of proteins that were different in at-risk participants compared with ACPA-negative controls, most of them elevated and enriched for cytokine and chemokine pathways. In short, the circulating blood of many ACPA-positive people looks as though the immune system is quietly on high alert.

Naïve T and B cells are not neutral bystanders, they are primed

One striking result is that even so-called naïve lymphocytes, the immune cells that have not yet seen their specific antigen, carried signatures of activation. The study detected transcriptional evidence that naïve CD4 T cells and naïve B cells in at-risk people were “prewired” to differentiate into effector cells. In functional tests, naïve B cells from at-risk participants produced more proinflammatory cytokines such as interleukin-6 after stimulation, and a subset showed early markers that make class switching to inflammatory IgG subclasses more likely. Together, these data indicate that the pool of cells normally reserved for first encounters with new pathogens is already skewed toward an inflammatory future.

An epigenetic hint: a putative IL21 enhancer appears in naïve CD4 T cells

Going beyond RNA, the team mapped chromatin accessibility to find regions of the genome more open to being read into RNA. They discovered a roughly 500-base-pair intronic region near the IL21 gene that was uniquely accessible in naïve CD4 T cells from at-risk individuals. That region overlaps an enhancer-like element previously seen in tonsillar T follicular helper cells and contains motifs for transcription factors such as BCL6 and STAT3 that push T cells toward a phenotype that helps B cells make high-affinity antibodies. In plain terms, the chromatin landscape in naïve T cells is permissive to becoming the very T cells that fuel pathogenic B cell responses.

Memory CD4 T cells with B-cell helper programs expand as disease approaches

As some at-risk participants progressed to clinical RA, the researchers observed expansion of a memory CD4 T cell subset with transcriptional and surface protein characteristics of T follicular helper–like or peripheral helper cells. These cells express classic B cell–help molecules, such as PD1, CXCR5 and ICOS, and their gene programs are tuned toward antigen receptor signaling and cytokine production that support B cell maturation and antibody production. The timing is important: this expansion happens in the months before clinical joint swelling, implicating these cells in the final push toward symptomatic disease.

Atypical and class-switched B cells shift toward inflammation

The B cell compartment showed a proinflammatory skew during progression, including expansion of “atypical” CD27-negative effector B cell clusters that express TBX21 (T-bet) and other markers associated with age-related or autoimmune B cell programs. Naïve B cells displayed increased sterile transcription at the IGHG3 locus, a molecular hint that these cells are poised to class switch to IgG3, an antibody subclass often enriched in inflammatory disease. Functionally, those naïve B cells made more IL-6 and RANKL after stimulation, both molecules that can promote inflammation and bone resorption.

Innate immune cells show a late, inflammatory surge at diagnosis

Monocyte populations, especially IL1B-expressing CD14 monocytes and interferon-stimulated gene–positive CD16 monocytes, rose in inflammatory gene expression at the time of clinical diagnosis. This suggests a model in which adaptive immune priming is present earlier, and innate inflammatory programs intensify as clinical disease becomes overt.

Autoantibody levels were not the only story

A common assumption is that rising ACPA titers foreshadow imminent joint disease. In this cohort, most converters did not have a clear longitudinal increase in circulating ACPA titers prior to diagnosis. Instead, cellular activation programs and protein signatures provided better clues to impending clinical RA than autoantibody titer alone. This suggests that antibody presence marks risk, but immune-cell behavior may predict timing.

Why this matters — implications for prevention, diagnosis, and therapy

These findings shift the window of “disease start” earlier into a preclinical phase marked by systemic inflammation and immune-cell reprogramming. That has three practical consequences.

First, biomarkers that combine proteomic, transcriptomic, and epigenetic signals could stratify which ACPA-positive people face the highest immediate risk of clinical RA, enabling monitoring over watchful waiting alone. The data imply that measuring only ACPA levels will miss the broader immune story.

Second, the discovery that naïve T cells are epigenetically poised to become B cell helpers suggests new prevention targets. Therapies that blunt T cell costimulation or block the specific differentiation pathways that produce TFH-like cells could interrupt the T-B cell feedback loop before joint damage occurs. Indeed, the scientists compared their preclinical gene signatures with profiles from abatacept trials, and the converter signatures were strongly modulated by abatacept responders but not by tumor necrosis factor inhibitor responders, suggesting a mechanistic match between the early T cell programs and CTLA4-Ig therapy.

Third, the work suggests a staged model of disease: early adaptive priming with later innate amplification. This helps explain why some clinical trials that target B cells, such as rituximab, delay onset in some participants but do not uniformly prevent RA. The timing and the target may both matter.

Caveats, limitations, and what we still do not know

No single study is definitive. The scientists note several limitations: the cohort is moderately sized and enriched for particular demographics, which raises questions about generalizability; peripheral blood may not capture all processes occurring inside the joint microenvironment; and the number of converters is relatively small for finely powered subgroup analyses. The scientists call for independent, ideally larger and more diverse, longitudinal cohorts and for paired synovial tissue studies to map local joint events alongside blood signatures.

Practical next steps for researchers and clinicians

For researchers, the dataset and interactive resource released by the team permit hypothesis testing across cell types, genes, and proteins, and they provide an open starting point to validate biomarkers in other cohorts.

For clinicians and trialists, these results argue for designing prevention trials that enroll participants based on a broader immunologic risk score rather than ACPA positivity alone, and for testing agents that modulate T cell costimulation or the T cell programs that enable B cell help, ideally at the moment when blood signatures indicate imminent conversion risk. Existing abatacept data are consistent with the idea that blocking T cell costimulation can reverse the very gene programs the scientists observe in converters.

Conclusion — a new picture of RA’s earliest stage

He and colleagues have provided a high-resolution portrait of the immune landscape that precedes clinical rheumatoid arthritis in people who are ACPA-positive but asymptomatic. Instead of a sudden ignition at the joint, RA appears to begin as a systemic process involving primed naïve lymphocytes, epigenetic permissiveness for B cell helper T cell differentiation, and a slow build toward innate inflammation. By moving our view of disease onset earlier and deeper into immune circuitry, this work opens realistic paths for better risk stratification, preemptive treatment trials, and ultimately prevention. The data and tools are publicly available for the community to test those paths.

Data and resources: Processed single-cell data are available from GEO, raw fastq files are deposited in dbGaP, and the authors host datasets and code on Dryad and GitHub for reproducibility and reuse.

The research was published in Science Translational Medicine on September 24, 2025.