AqF026 is a pharmacologic agonist of the water channel aquaporin-1

Abstract

Aquaporin-1 (AQP1) facilitates the osmotic transport of water across the capillary endothelium, among other cell types, and thereby has a substantial role in ultrafiltration during peritoneal dialysis. At present, pharmacologic agents that enhance AQP1-mediated water transport, which would be expected to increase the efficiency of peritoneal dialysis, are not available. Here, we describe AqF026, an aquaporin agonist that is a chemical derivative of the arylsulfonamide compound furosemide. In the Xenopus laevis oocyte system, extracellular AqF026 potentiated the channel activity of human AQP1 by >20% but had no effect on channel activity of AQP4. We found that the intracellular binding site for AQP1 involves loop D, a region associated with channel gating. In a mouse model of peritoneal dialysis, AqF026 enhanced the osmotic transport of water across the peritoneal membrane but did not affect the osmotic gradient, the transport of small solutes, or the localization and expression of AQP1 on the plasma membrane. Furthermore, AqF026 did not potentiate water transport in Aqp1-null mice, suggesting that indirect mechanisms involving other channels or transporters were unlikely. Last, in a mouse gastric antrum preparation, AqF026 did not affect the Na-K-Cl cotransporter NKCC1. In summary, AqF026 directly and specifically potentiates AQP1-mediated water transport, suggesting that it deserves additional investigation for applications such as peritoneal dialysis or clinical situations associated with defective water handling.

Figure 1.

Potentiating effect of AqF026 on water channel activity of AQP1 and AQP4 expressed in Xenopus laevis oocytes. (A) Change in volume (V) due to osmotic swelling, standardized to the initial volume (V₀) and plotted as a function of time in 50% hypotonic saline, for AQP1-expressing and -nonexpressing control (cont) oocytes. Oocytes were preincubated in 10 μM AqF026 or with DMSO alone (untr) as a vehicle control. Data are mean ± SEM for all oocytes tested in a single experimental day; n values are indicated in italics. (B) Histogram of compiled data showing maximal potentiation of AQP1 near 5 μM AqF026 and no potentiation of AQP4 at doses <50 μM. (C) Dose-response relationships for AqF026-mediated potentiation of AQP1 and AQP4 water channel activities, with an estimated EC₅₀ value of 3.3 μM and a Hill coefficient of 1.8 for the stimulatory component for AQP1 (fit as the sum of two dose-response curves, one stimulatory and one inhibitory, using GraphPad Prism).

Figure 2.

Ligand docking of AqF026 on AQP1: involvement of loop D domain residues and specificity of action using an NKCC1-sensitive bioassay. The identification of residues associated with the proposed intracellular binding pocket are based on the crystal structure of bovine AQP1. (A) Surface rendition of the tetramer with the docked pose of AqF026 in a ribbon structure of one AQP1 subunit. (B) Chemical structure of AqF026 and the parent compound furosemide. (C) A detailed view of the amino acid side chains relevant to the docked pose. The highlighted residue (D165*) indicates an amino acid contribution from a neighboring subunit. (D) Relative swelling rates were standardized to the untreated AQP1 wild-type or mutant channel responses within the same batches of oocytes, and data were compiled from at least three different batches for each group. Data are mean ± SEM; n values are shown in italics above the x axis. Asterisks indicate a significantly different effect for the comparison of 0 and 10 μM AqF026 for the same construct (*P<0.05; *** P<0.001). There was no significant difference between wild type and G165P at 10 μM (not indicated). (E) Confocal images of permeabilized immunolabeled oocytes expressing AQP1 wild-type or mutants, showing channel protein localization in plasma membranes (arrows). (F) Intracellular recording from a mouse gastric antrum preparation, continuously measured from a single circular smooth muscle cell. Traces show representative segments from the sequence: first confirming stable spontaneous slow-wave activity (initial); after treatment with a methylated furosemide compound (AqF022, 10 μM), which caused depolarization and a concomitant decline in slow-wave amplitude; after reversal of the effect by washout (recovery); and during the subsequent lack of effect of AqF026 (10 μM) over an extended period (30 minutes). Similar results were seen in three replicate cells.

Figure 3.

AqF026 specifically enhances AQP1-mediated osmotic water transport in vivo. (A) Wild-type Aqp1^+/+ mice treated with AqF026 have a significant, dose-dependent increase in net ultrafiltration (UF) across the peritoneal membrane. The maximal response is observed for a concentration of 15 µM, with an estimated EC₅₀ value of 4.2 µM (inset). Aqp1^−/− mice, characterized by a 60% reduction in ultrafiltration at baseline, show no potentiation of the ultrafiltration after treatment with 15 μM AqF026. Data are mean ± SEM; n=6 for each AqF026 concentration except for 0.75 µM (n=4). Net ultrafiltration rates were standardized to body weight and compared with rates in vehicle-treated mice (open bar). (B) Time course performed in wild-type mice shows that the effect of AqF026 (15 µM) on net ultrafiltration (bars) and initial ultrafiltration rate (triangles) is maximal 120–150 minutes after intravenous injection. Data are mean ± SEM; n=4 for each time point. Open bar corresponds to vehicle-treated mice. (C) Treatment of Aqp1^+/+ mice with 15 µM AqF026 results in increased intraperitoneal (IP) volume over time (P=0.02 between the AqF026 and vehicle curves), with significant differences at 30, 60, and 90 minutes in AqF026-treated (black triangles) compared with vehicle-treated (open squares) animals (P<0.01, P<0.05, and P<0.05 respectively; n=10 in each group). (D) Initial ultrafiltration rates, taken as an index of AQP1-mediated water transport during the first part of the dwell, are significantly increased in Aqp1^+/+ mice treated with 15 μM AqF026 versus vehicle (P<0.001, n=10 in each group). In contrast, AqF026 has no effect on the initial ultrafiltration rates in Aqp1^−/− mice (AqF026, n=4 and vehicle-treated, n=5). (E) Treatment of wild-type mice with AqF026 (15μM) induces a significant increase in sodium sieving (D/D₀ sodium at 30 min) compared with vehicle-treated animals (P<0.05; n=10 in each group). (F–H) The dialysate-to-plasma ratio of osmolality at 30 minutes (D/P osm) (F), the dialysate-to-plasma ratio of urea (D/P urea) (G), and the progressive removal of glucose from the dialysate (D/D₀ glucose) (H) were similar in mice treated with 15μM AqF026 versus vehicle (n=12 pairs of Aqp1^+/+ mice).