The fibers at the center of the study are known as curli, proteinaceous amyloids secreted by members of the phylum Pseudomonadota, a broad bacterial group that includes Salmonella enterica serovar Typhimurium and Escherichia coli. When these bacteria form biofilms, dense and resilient communities attached to surfaces, they weave curli fibers through an extracellular matrix that also traps nucleic acids. The result is a composite structure in which amyloid protein and DNA are physically intertwined. This is a biologically provocative combination, because amyloids are famously resistant to degradation and because DNA, when presented to the immune system in the right context, can activate innate immune pathways that push the body toward antibody production. Earlier work by the same team had demonstrated that complexes of curli and DNA could elicit autoantibodies against double-stranded DNA and chromatin in mice, hallmarks of the autoimmune response seen in lupus patients.

In the new study, the researchers set out to ask a broader question: what happens to the autoimmune landscape when animals are exposed to curli, or infected with Salmonella, which naturally produces curli-rich biofilms? They turned to a multiplex autoantibody assay, a technology that allows dozens of antibody specificities to be measured simultaneously in a small volume of serum. When they compared mice treated with curli against control animals injected with phosphate-buffered saline, a clear signal emerged. Both ordinary C57BL/6 laboratory mice and lupus-prone NZBxW/F1 mice, a strain that spontaneously develops a lupus-like disease, showed elevated levels of antibodies against C1q after curli exposure. This was not a subtle effect confined to a single genetic background; it appeared in healthy mice and in mice destined to develop autoimmunity alike.

To confirm and extend the multiplex results, the team performed enzyme-linked immunosorbent assays, or ELISAs, specific for anti-C1q antibodies. The confirmation held. C57BL/6 mice developed anti-C1q antibodies after treatment with curli-DNA complexes extracted from E. coli biofilms, after treatment with similar complexes derived from Salmonella biofilms, and, notably, after live infection with Salmonella Typhimurium itself. This last result is perhaps the most consequential, because it ties the generation of a lupus-associated autoantibody to a real infectious event rather than to the artificial administration of an isolated bacterial product. It suggests that the ordinary experience of infection with biofilm-forming bacteria may be sufficient to nudge the immune system toward the production of antibodies that recognize components of the body’s own complement machinery.

C1q is not an arbitrary target. It is the initiating molecule of the classical complement pathway, a cascade of proteins that helps clear immune complexes, dead cells, and pathogens. In lupus, particularly in lupus nephritis, C1q becomes deposited in the glomeruli, the filtering units of the kidney, where it is thought to be stripped from immune complexes and exposed on the surfaces of apoptotic debris. Anti-C1q antibodies that bind to this deposited protein are strongly correlated with active nephritis and are among the best predictors of kidney involvement in lupus patients. The discovery that bacterial amyloids can elicit these antibodies therefore provides a plausible mechanistic bridge between infection, dysbiosis, and renal disease, three phenomena that clinicians have long observed to travel together in lupus patients without a satisfying explanation of how they connect.

The researchers then probed how timing and genetic susceptibility shaped the response. In young NZBxW/F1 mice, before the spontaneous onset of their lupus-like disease, injection of curli produced significantly higher levels of both anti-dsDNA and anti-C1q antibodies compared with controls. Yet in older mice, after disease onset, curli made no measurable difference, because autoantibody levels were already elevated across the board. This pattern suggests that bacterial amyloid exposure may matter most early in the disease process, when it can help initiate or amplify the autoimmune response before it becomes self-sustaining. Once the autoimmune machinery is running at full speed, an additional push from microbial products may be lost in the noise of an already hyperactive immune system.

A different picture emerged in MRL/lpr mice, another lupus-prone strain characterized by a mutation that causes profound lymphoproliferation and aggressive disease. In these animals, curli treatment significantly exacerbated anti-C1q antibody production regardless of the age of the mice or the stage of their disease. The contrast between the two strains is instructive. It implies that the impact of bacterial amyloid exposure on autoantibody production depends on the immunological context: in a strain where the autoimmune response is still building, curli acts as an accelerant during the early phase, whereas in a strain with a more explosive and pervasive autoimmune program, curli can intensify specific arms of the response at any point. Either way, the microbial signal is not passive; it actively shapes the autoantibody repertoire.

What about the clinical consequences? Interestingly, curli treatment did not significantly change proteinuria, the leakage of protein into the urine that serves as a standard readout of kidney damage, in either lupus model. But a more direct examination of kidney tissue told a different story. In MRL/lpr mice, curli treatment significantly increased the deposition of C1q in the glomeruli, the very site where anti-C1q antibodies are believed to inflict damage in human lupus nephritis. The combination of more C1q deposited in the kidney and more antibodies circulating against C1q paints a coherent picture of a mechanism by which bacterial exposure could worsen renal inflammation, even if the effect does not immediately register in a crude functional measure like proteinuria over the time course of the experiment.

The study also connects to a growing body of evidence implicating the microbiome in lupus. Patients with SLE and lupus nephritis frequently show dysbiosis, a disturbance of the normal microbial community, often marked by an overgrowth of Pseudomonadota, the same phylum that produces curli. Infections are also frequent in lupus patients and are well documented to associate with disease flares, increasing morbidity and mortality. The new findings suggest that these observations may not be separate curiosities but different faces of a single mechanism: biofilm-forming bacteria, whether flourishing in the gut during dysbiosis or invading during infection, expose the immune system to amyloid-DNA complexes that can elicit antibodies against DNA and against C1q, thereby feeding the autoimmune process that defines the disease.

Of course, mice are not people, and the leap from curli-treated laboratory animals to human lupus patients requires caution. The researchers did not demonstrate curli-driven autoantibody production in humans in this study, and the precise contribution of bacterial amyloids to the initiation of human disease remains to be established. Even so, the work opens concrete avenues for investigation. If curli and related amyloids from the microbiome do contribute to lupus in patients, then measuring exposure to these structures, monitoring anti-C1q responses in the context of infections or dysbiotic episodes, and ultimately exploring interventions that reduce biofilm burden or block the immune recognition of amyloid-DNA complexes could become meaningful strategies. For a disease in which kidney involvement remains a leading cause of suffering, a mechanism that links infection and microbial ecology to the production of nephritis-associated autoantibodies is a finding worth taking seriously. The study, published in PLOS Pathogens, uncovers a novel route by which the microbial world may help ignite and sustain the systemic autoimmunity of lupus.

Subject of Research: The induction of anti-C1q and anti-dsDNA autoantibodies by Salmonella infection and bacterial biofilm curli-DNA complexes in mouse models of systemic lupus erythematosus.

Article Title: Salmonella infection and biofilms drive anti-C1q and anti-dsDNA autoantibodies in systemic autoimmunity

Article References: Grando, K., Elkins, M., Nicastro, L., Bessho, S., Olubajo, S., Kowal, A., Tam, V., Caricchio, R., Pisetsky, D., Silverman, G. J., & Tükel, Ç. (2026). Salmonella infection and biofilms drive anti-C1q and anti-dsDNA autoantibodies in systemic autoimmunity. PLOS Pathogens, 22(10), e1014666. https://doi.org/10.1371/journal.ppat.1014666

Image Credits: AI Generated

DOI: 10.1371/journal.ppat.1014666

Keywords: systemic lupus erythematosus, Salmonella, biofilms, curli, anti-C1q antibodies, anti-dsDNA antibodies, lupus nephritis, dysbiosis, autoimmunity, PLOS Pathogens, complement C1q, mouse models