Found in a rare disease. It did not stay there.

It began with a transplanted kidney

Focal segmental glomerulosclerosis can return in a transplanted kidney within days. The new kidney is healthy, so whatever is destroying it arrived with the recipient.

The cell it destroys is the podocyte, whose interlocking foot processes form the slits plasma is filtered through. Primary focal segmental glomerulosclerosis returns in 30% to 80% of transplanted kidneys, and in someone who has already lost one graft to recurrence the risk in the next is 80% to 100% (Shoji, Nephron 2020). Something in the blood does this, and the field looked for it for decades without finding it.

suPAR was proposed as that factor in 2011. It was raised in roughly two thirds of people with primary focal segmental glomerulosclerosis, highest before transplantation in those who went on to recur, and lowered by plasmapheresis. In mice, sustained suPAR expression produced protein in the urine and flattened the podocyte's foot processes; a mutant unable to bind αvβ3 integrin produced neither, and an anti-uPAR antibody prevented the injury (Wei, Nat Med 2011). That built on an earlier finding that uPAR signaling in the podocyte itself activates αvβ3 and opens the filter (Wei, Nat Med 2008).

It did not stay there

Whether suPAR causes that one rare disease is still argued. What is no longer in question is that it belongs to kidney disease generally, and the evidence for that came from people who did not have focal segmental glomerulosclerosis at all.

We measured it in 3,683 participants of the Emory Cardiovascular Biobank and followed their kidney function. Filtration fell by 4.2 mL/min/1.73 m² per year in the highest quartile of suPAR against 0.9 in the lowest, and among the 1,335 who did not have chronic kidney disease at the outset, the highest quartile was three times as likely to develop it. The steepest decline of all was in the 921 participants whose filtration was entirely normal to begin with, which is the group conventional testing has nothing to say about (Hayek, N Engl J Med 2015).

The finding has since held in populations that share almost nothing else. Among 898 European children, five-year survival free of kidney failure was 64.5% in the lowest quartile of suPAR and 35.9% in the highest (Schaefer, JAMA Pediatr 2017). Among 649 adults with autosomal dominant polycystic kidney disease, 68% of the highest tertile reached stage 3 chronic kidney disease within three years, against 22% of the lowest (Hayek, J Am Soc Nephrol 2019). It predicts progression in 2,391 Chinese patients with chronic kidney disease (Lv, Nephrol Dial Transplant 2020), and in Black Americans with chronic kidney disease it adds risk on top of filtration measured directly rather than estimated (Luo, Clin J Am Soc Nephrol 2018). Different diseases, different ancestries, the same signal.

Made in the marrow, felt in the filter

suPAR is not a kidney protein. It is made by immune cells, and the kidney is one of the places it lands.

In mice, immature myeloid cells in the bone marrow account for the pathological levels, and transferring those cells into healthy animals transfers the proteinuria (Hahm, Nat Med 2017). In people, bone marrow from patients with chronic kidney disease carries raised TNFα and suPAR alongside inflammatory monocytic cells, and myeloid cells reprogrammed by TNFα in culture secrete more suPAR, whose secretions disorganize the podocyte's skeleton (Spear, J Am Soc Nephrol 2026).

This is also why a level is not a measure of kidney function. suPAR is not filtered by the kidney: regional sampling in people finds only modest renal extraction, and the heart does the same (Chew-Harris, Clin Biochem 2019), while a dialysis session that removes 58% of β2-microglobulin leaves suPAR unchanged (Kampmann, Scand J Clin Lab Invest 2024). What a level reports is the inflammatory state producing it (Hayek, Circ Res 2026).

At the filter, circulating suPAR engages αvβ3 integrin together with RAGE on podocytes, cells that never made the receptor, and that pairing is required for the signaling that flattens their foot processes (Kim, Biochim Biophys Acta Mol Basis Dis 2021). Not every form of the protein does the same thing. A second uPAR isoform forms a dimer and causes severe scarring in mice through β3 integrin and c-Src, which a Src inhibitor blunts (Wei, J Clin Invest 2019); the D2D3 fragment injures the kidney and the insulin-producing beta cell both, and an anti-uPAR antibody restores beta cell mass (Zhu, Sci Transl Med 2023).

A genetic risk that turned out to be conditional

Two variants in the APOL1 gene raise the risk of kidney failure sharply in people of recent African ancestry. Most carriers never develop kidney disease, so something else has to be present.

In two large, unrelated cohorts, the decline in kidney function associated with the APOL1 risk variants depended on the person's suPAR level: the genetic risk was attenuated when suPAR was low and strengthened when it was high. The risk proteins bind suPAR-activated αvβ3 integrin more tightly than the reference protein does, and in mice they caused proteinuria only when suPAR was present (Hayek, Nat Med 2017). The variants are not simply toxic. They answer a signal the immune system is already sending.

A separate cohort points the same way: in Black Americans with chronic kidney disease, suPAR predicted worsening proteinuria only among those carrying two APOL1 risk alleles (Luo, Clin J Am Soc Nephrol 2018). The implication is not modest. A genetic risk that had looked fixed depends on a circulating protein, and circulating proteins can be lowered.

When the insult has a date

Much of the kidney injury that happens in a hospital is scheduled: a contrast injection, a bypass run, an admission to intensive care. A level drawn beforehand identifies who is at risk.

We measured suPAR before coronary angiography in 3,827 patients, before cardiac surgery in 250, and on admission to intensive care in 692. The highest quartile carried 2.66 times the adjusted odds of acute kidney injury within seven days. The same study went further than association: mice overexpressing suPAR sustained worse injury from contrast material, human kidney tubular cells exposed to it showed oxidative stress, and an anti-uPAR antibody prevented both (Hayek, N Engl J Med 2020). Preprocedural risk is knowable, and in an animal it is preventable.

Infection makes the point more sharply. Among 352 patients admitted with COVID-19, acute kidney injury occurred in 6% of the lowest suPAR tertile and 46% of the highest, and nobody below 4.60 ng/mL needed dialysis (Azam, J Am Soc Nephrol 2020). In sepsis, mice lacking suPAR were protected and mice overexpressing it did worse (Nusshag, JCI Insight 2023). Some of the proteinuria of a severe viral illness is the immune response itself, acting through suPAR on podocyte integrins (Wei, Nat Commun 2023).

What it would take to treat it

Every mechanism above has a matching intervention that works in an animal. None of them is yet a treatment for a person.

The anti-uPAR antibody is the common thread: it prevents suPAR-driven proteinuria in mice, attenuates contrast-induced kidney injury, and restores beta cell mass (Wei, Nat Med 2011; Hayek, N Engl J Med 2020; Zhu, Sci Transl Med 2023). Plasmapheresis lowers suPAR in people, but it removes a great deal besides suPAR, so it cannot establish what is doing the work (Wei, Nat Med 2011). The one trial built around this biology gave rituximab to nine adults with treatment-resistant disease selected for a high suPAR, and proteinuria did not improve (Hladunewich, Kidney Int Rep 2022). Rituximab depletes B cells and does not lower suPAR, so the target itself has still not been tested in people by an agent that hits it.

What has changed is the reason to try. suPAR is no longer only a marker that sorts patients by risk. It has a receptor complex on the podocyte, a genetic interaction that explains who is susceptible, a source in the bone marrow, and an effect that can be blocked in an animal at three separate points. The kidney is also not the only organ it reaches, which is the case for treating cardiovascular, kidney and metabolic disease as one problem rather than three (Reiser, J Clin Invest 2026; Hayek, Circ Res 2026).