Background and hypothesis: Sodium-glucose cotransporter 2 (SGLT-2) inhibitors (SGLT2i) improve kidney and cardiovascular outcomes in patients with diabetes and chronic kidney disease (CKD). Since complement system precursors are aberrantly filtered and activated in conditions with albuminuria, it was hypothesized that SGLT2i lower plasma collectins and attenuate intratubular complement activation and deposition in patients with CKD.

Methods: Patient plasma and urine samples were analyzed from three randomized, blinded, cross-over intervention studies where patients with type 2 diabetes mellitus (DM) and CKD (DM-CKD, n=17), DM without CKD (DM, n=16), and CKD without DM (CKD, n=16) were given empagliflozin (10 mg/day) and placebo for four weeks with a washout of at least two weeks in-between. ELISA was used to determine concentrations of collectin kidney 1 (CL-K1), collectin liver 1 (CL-L1), mannose-binding lectin (MBL), MBL-associated serine protease (MASP-2), anaphylatoxins C3a and C5a, stable C3 split product C3dg and the membrane attack complex (sC5b-9) in urine and plasma. Apical deposition of membrane-bound C5b-9 was determined in isolated extracellular vesicles (EVs) from urine by western blotting (n=5). A paired two-tailed t-test was used to determine statistical significance between the treatments in each group. Non-normal distributed values were reported as median [interquartile range]. The studies were approved by the Ethical Committee of the Central Jutland Region and the Danish Health and Medicine Authority. The collection of samples was performed in accordance with the Helsinki Declaration, the EU Directive on Good Clinical Practice (GCP), and International Conference of Harmonization (ICH-GCP) guidelines and all participants gave written informed consent before inclusions.

Results: Empagliflozin induced no change in plasma levels of CL-K1, CL-L1, MBL, MASP-2, C3a, C5a, and C5b-9 while C3dg levels increased (18%, 114.1 [91-139] vs. 131.9 [98-157] units/mL, p=0.03) in DM-CKD. Collectins were not detectable in protease-inhibited spot urine or in ex vivo up-concentrated urine. Empagliflozin decreased urine C3a/creatinine ratio in the DM (32.5%, 5.7 [2.6-15] vs. 4.1 [1.8-6.4] ng/µmol, p=0.01) and DM-CKD groups (58.3%, 5.1 [3.6-238] vs. 4.7 [2.0-105] units/µmol, p=0.012), whereas urine C5a- and C3dg/creatinine ratio did not change. Urine sC5b-9/creatinine decreased significantly (45%, 0.9 [0.1-17] vs. 0.4 [0.1-12] units/µmol, p=0.02) after empagliflozin but only in the DM-CKD group. This was recapitulated in urine EVs by immunoblotting for C5b-9. Empagliflozin reduced the urine albumin/creatinine ratio in the DM-CKD group (24.5%, 201 [62-1130] vs. 164 [44-719] mg/g, p=0.02).

Conclusion: While SGLT2i do not lower plasma concentrations of collectins and complement activation products, they significantly reduce indices of intrarenal complement activation and membrane deposition in patients with DM and DM-CKD. Thus, SGLT2 inhibitors may protect kidneys in diabetic patients with CKD by reducing complement-induced tubular injury.