Diabetic kidney disease (DKD) occurs in up to 20%–40% of patients with diabetes and has become the leading cause of chronic kidney disease (CKD), renal failure, and premature death (1). The classic clinical phenotype of DKD consists of persistently increased albuminuria and a progressive decline in glomerular filtration rate (GFR) (2). However, alternative trajectories of kidney function in patients with diabetes have been reported and characterized by a rapid decline in GFR, regression in albuminuria, or absence of albuminuria or proteinuria (3). The annual median estimated GFR (eGFR) varied from -4.0 to -1.5 ml/min/1.73m² in patients with diabetes (4). Those with an eGFR slope ≤ -5 ml/min/1.73m²/year are termed as having rapid kidney function decline (5). A higher rate of eGFR decline was considered to be associated with a higher risk of subsequent renal failure and cardiovascular comorbidities (6).

Observational studies suggested that albuminuria or macroalbuminuria may not indicate a rapid decline in eGFR (7). Although kidney biopsies are not typically required for the diagnosis of DKD, a strong relationship has been observed between histological findings and clinical presentations and progressions (8). A loss of kidney tubules correlates with the decline in eGFR in advanced DKD (9). Ferroptosis is an iron-dependent regulated cell death that is driven by robust cellular lipid oxidation and is inhibited by thiol-dependent peroxidase (10). Tubular cells have active iron reabsorption, a high lipid metabolic rate, and concentrated redox-active compound perturbations, which constitute the main pillars of ferroptosis (11, 12). Our previous data revealed beneficial effects on kidney tubular protection in diabetes by suppressing ferroptosis (13). It was reported that upregulation of the pro-ferroptotic protein long-chain fatty acid CoA ligase 4 (ACSL4) appeared in tubular cells in various human kidney diseases (14). The tubular expression of ACSL4 has been observed in mouse models and patients with DKD, and it was negatively correlated with the patients’ eGFR (15). However, the previous study group consisted of various etiologies and lacked specificity in DKD. Moreover, the association between the tubular expression of ACSL4 and kidney outcomes of patients with DKD is unknown.

Therefore, in this retrospective cohort study, we enrolled patients with biopsy-proven DKD to investigate the association between the tubular expression of ACSL4 and a rapid decline in kidney function.

In this study, immunostaining showed that ACSL4 was mainly expressed in tubular epithelial cells in kidney biopsy specimens from human subjects. The increased expression levels of ACSL4 in kidney tubules from patients with DKD were further positively correlated with proteinuria at the time of the renal biopsy. Multivariate Cox proportional hazard regression models adjusted for sex, baseline age, blood pressure, HbA1c, eGFR, and proteinuria showed that patients with a high baseline tubular expression level of ACSL4 were at high risk for rapid kidney function decline when ACSL4 was modeled as both a continuous variable (1-SD increment in log2-transformed IOD/area of tubular ACSL4) and a categorical variable (the lowest tertile as reference).

Common risk factors for the rapid progression of DKD include high blood glucose, high blood pressure, proteinuria, and low GFR category (20). It is assumed that a rapid decline in kidney function has likely become less common since multifactorial interventions have become available in clinical settings. Notably, our previous study indicated that sodium-glucose cotransporter 2 (SGLT2) inhibitors and renin-angiotensin-aldosterone system inhibitors were not routinely used in more than 66% and 77% of the DKD patients we followed, respectively (21). Nevertheless, 26/72 (36.11%) patients with DKD progressed rapidly, as demonstrated by an eGFR slope < -5 ml/min/1.73m²/year. An increase in tubular expression levels of ACSL4 evaluated at the time of the renal biopsy helped identify patients with a high risk of rapid kidney function decline independent of the baseline blood glucose, blood pressure, proteinuria, and eGFR, suggesting a new approach for precision intervention. Additionally, this study may validate tubular ACSL4 as an early marker with the advantage of pre-dating the functional decline and other risk factors.

Previous studies found tubular cell ferroptosis in models of diabetes and suggested that inhibition of ferroptosis relieved the development and progression of DKD (13, 22–24). Increased ACSL4 facilitates the availability of lipid substrates and favors ferroptosis. However, the presence of ACSL4 in the control subjects does not necessarily indicate ferroptosis as ACSL4 itself is involved in lipid biosynthesis. Moreover, a deficiency of ACSL4 prevented ferroptosis but favored the occurrence of necroptosis (25). The pathologic mechanisms of increased ACSL4 in tubular epithelial cell death and progression of DKD should be investigated in further studies.

Consistent with our study, a previous study found that decreased tubular expression of glutathione peroxidase 4 (GPX4) predicted DKD progression (26). GPX4 is an intracellular antioxidant enzyme that protects cells against lipid peroxidation and therefore prohibits ferroptosis (27). In animal models or cultured cells, ferroptosis is always evaluated by ACSL4 upregulation, GPX4 reduction, ferrous iron overload, and lipid peroxidation aggregation. It is plausible that the increased expression of ACSL4 may be an appropriate pathologic biomarker for clinical applications.

Proteinuria typically indicates dysfunction in the glomerular filtration barrier, while ACLS4 expression in tubules suggests injuries of tubular cells. The relationship between proteinuria and ASCL4 may indicate an interplay between tubules and glomeruli (28) underlying the progression of DKD and led us to consider tubuloglomerular feedback mechanisms. Furthermore, significantly reduced proteinuria and mitigated glomerular hypertrophy were found by suppressing ferroptosis in a diabetic model (13). However, it was not easily explained by directly blocking adverse glomerular hemodynamics, such as with SGLT2 inhibitors. Further studies are needed.

It cannot be definitively stated that a person with diabetes and CKD has DKD unless the person has a kidney biopsy; however, in most cases, a kidney biopsy was not recommended as other possible diagnoses would not change treatment (29). A kidney biopsy is performed due to the presence of at least one of the following indications (17): rapid onset of proteinuria, presence of active urinary sediment (e.g., hematuria), rapid decrease in renal function, absence of retinopathy, resistant hypertension, and suspicion of other nephropathies secondary to systemic disease. Therefore, kidney biopsy samples from patients in the DKD stage I or IV were scant due to the absence of indications or the presence of contraindications. We observed a positive correlation between tubular ACSL4 and proteinuria, which probably indicates a tubuloglomerular interplay as we mentioned above; however, the relationship between tubular expression of ACSL4 and diabetic nephropathy defined by the Renal Pathology Society as well as proteinuria, which both primarily describe glomerular lesions, remains obscure.

The current study has some strengths. We identified tubular expression levels of ACSL4 as a novel biomarker for rapid kidney function decline in subjects with DKD. There were some limitations of our study. The sample size was limited and most of the patients enrolled were at the DKD stages II and III (16). The role of ACSL4 in the progression of DKD remains obscure and whether ACSL4 contributes to CKD with other etiologies needs further investigation.

In conclusion, this study showed that a significantly increased expression of ACSL4 in kidney tubules was associated with more severe proteinuria and a higher risk of rapid kidney function decline as evaluated by eGFR slope in patients with DKD.