Insulin and IGF-1 activate K_ir4.1/5.1 channels in cortical collecting duct principal cells to control basolateral membrane voltage

Reagents and animals.

All chemicals and materials were from Sigma (St. Louis, MO), VWR (Radnor, PA), and Tocris (Ellisville, MO) unless noted otherwise and were at least of reagent grade. Animal use and welfare adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals following protocols reviewed and approved by the Animal Care and Use Committees of the University of Texas Health Science Center at Houston and the Medical College of Wisconsin. For experiments, male C57BL/6J mice (Charles River Laboratories, Wilmington, MA), 6–10 wk old, were used. Animals were maintained on a standard rodent regimen and had free access to tap water.

Tissue isolation.

The procedure for isolation of the CCD suitable for electrophysiology is a modification from the protocols described previously (20–22, 44). Mice were euthanized by CO₂ administration followed by cervical dislocation, and the kidneys were removed immediately. Kidneys were cut into thin slices (<1 mm) with slices placed into ice-cold physiological saline solution (PSS) containing (in mM) 150 NaCl, 5 KCl, 1 CaCl₂, 2 MgCl₂, 5 glucose, and 10 HEPES (pH 7.35). Straight cortical-to-medullary sectors, containing ∼30–50 renal tubules, were isolated by microdissection using watchmaker forceps under a stereomicroscope. Isolated sectors were further incubated in PSS containing 0.8 mg/ml collagenase type I (Alfa Aesar, Ward Hill, MA) and 5 mg/ml of dispase II (Roche Diagnostics, Mannheim, Germany) for 20 min at 37°C followed by extensive washout with PSS. Individual CCDs were visually identified by their morphological features (pale color, coarse surface, and, in some cases, bifurcations) and were mechanically isolated from the sectors by microdissection. Isolated CCDs were attached to a 5 × 5-mm coverglass coated with poly-l-lysine. A coverglass containing a CCD was placed in a perfusion chamber mounted on an inverted Nikon Eclipse Ti microscope and perfused with PSS at room temperature. The tubules were used within 1–2 h after isolation.

Immunohistochemistry.

Kidneys were fixed in 10% formalin and processed for paraffin embedding. Kidney sections (4 μm) were double-labeled with rabbit anti-K_ir5.1 (SAB4501636, Sigma) and aquaporin-2 water channel (AQP2; sc-28629, Santa Cruz Biotechnology, Dallas, TX) or rabbit anti-K_ir4.1 (APC-035, Alomone Labs) and thiazide-sensitive Na⁺-Cl⁻ cotransporter (NCC; AB3553, Millipore, Billerica, MA) antibodies. Bindings were revealed with Alexa Fluor 488 or 643 with goat anti-rabbit biotinylated IgG (Molecular Probes, Waltham, MA). Immunostaining was performed in tissues from at least four different kidneys. All tissue sections were examined by TCS SP5 confocal laser-scanning microscopy (Leica, Buffalo Grove, IL).

Whole cell currents and membrane potential.

Whole cell currents in CCD cells were measured under voltage-clamp conditions in the perforated-patch mode with GΩ seals formed on the basolateral membrane. All patch-clamp data were acquired with an Axopatch 200B (Molecular Devices, Sunnyvale, CA) patch-clamp amplifier interfaced via a Digidata 1440 (Molecular Devices) to a computer running pClamp 10.4 (Molecular Devices). The bath solution was (in mM) 150 NaCl, 5 KCl, 1 CaCl₂, 2 MgCl₂, 5 glucose, and 10 HEPES (pH 7.35). Freshly made amphotericin-B (400 μM, Enzo Life Sciences, Farmingdale, NY) was dissolved in the pipette solution containing 150 mM K acetate, 5 mM KCl, 2 mM MgCl₂, and 10 mM HEPES (pH 7.35) by ultrasonication. Electrical recordings were made once the access resistance from the pipette to the cell interior fell to <15 MΩ, usually 5–10 min after achievement of a pipette-to-membrane seal resistance of 5–10 GΩ. Capacity of individual principal cells was on average 15 pF and was manually compensated. Real-time changes in membrane voltage in CCD cells were studied under current-clamp mode using the perforated-patch technique as was described previously (43, 45).

Single-channel recordings.

Single-channel activity of K_ir4.1/5.1 in CCD cells was determined in cell-attached patches on the basolateral membrane made under voltage-clamp conditions. Recording pipettes had resistances of 8–10 MΩ. Bath and pipette solutions were (in mM) 150 NaCl, 5 mM KCl, 1 CaCl₂, 2 MgCl₂, 5 glucose, and 10 HEPES (pH 7.35) and 150 mM KCl, 2 mM MgCl₂, and 10 mM HEPES (pH 7.35), respectively. In the cell-attached configuration, the actual voltage applied to a membrane patch (V_patch) is a sum of the pipette voltage and the resting basolateral membrane potential (V_basolateral) of principal cells, which is close to −70 mV for principal and −15 mV for intercalated cells (see Fig. 1).

For paired patch-clamp experiments, pipette voltage was −V_p = −40 mV. Current recordings were made in a permanently perfused bath (1.5 ml/min). Drug administration times are shown with bars on the top of representative traces. All applied reagents were prepared as 1,000× stock solutions in water or dimethyl sulfoxide (DMSO) as instructed. There were no changes in single-channel activity and macroscopic current amplitude during vehicle application. For each experimental condition, CCDs from at least three different mice were assayed. Currents were low-pass filtered at 1 kHz with an eight-pole Bessel filter (Warner Instruments, Hamden, CT). Events were inspected visually before acceptance. In paired experiments, K_ir4.1/5.1 activity was analyzed over a span of 60–120 s for each experimental condition after reaching a new steady state in response to a treatment. Channel activity (NP_o) and P_o were assessed using Clampfit 10.4. For calculating P_o in paired experiments, N was fixed as the greatest number of active channels observed in control or experimental conditions. For representation, current traces were filtered at 200 Hz and corrected for a slow baseline drift as necessary.

Data analysis.

All summarized data are reported as means ± SE. In paired experiments, data from before and after treatment were compared using the paired t-test. Data from unpaired experiments were compared with a Student's (two-tailed) t-test or a one-way ANOVA as appropriate. P < 0.05 was considered significant.

Basolateral K⁺ conductance via K_ir4.1 and K_ir4.1/5.1 is the major contributor to the whole cell current recorded from the basolateral side in CCD principal cells.

The CCD is known to contain principal (∼70% from total cell population) and intercalated (the remaining 30%) cells, which are presumably not electrically coupled (41). Indeed, we detected two independent cell populations maintaining resting membrane potentials of −70.2 ± 1.9 mV (n = 12) and −14.3 ± 2.5 mV (n = 7), as assessed in current clamp mode in patches on the basolateral membrane of freshly isolated CCDs (Fig. 1A). These values are close to the estimated reversal potentials for K⁺ (first group) and Cl⁻ (second group) assuming physiological distribution of the ions. These groups are consistent with principal and intercalated cells known to have predominantly potassium and chloride basolateral conductance, respectively (23). In further studies, we focused exclusively on principal cells. A typical whole cell current in response to a voltage-step protocol is shown in Fig. 1B. As is clear from the respective current-voltage relationship (I–V) in Fig. 1C, whole cell current demonstrates a noticeable rectification at negative voltages. Extracellular application of 30 μM Ba²⁺ and equimolar substitution of intracellular K⁺ by Cs⁺ to block K⁺ channels drastically reduce the amplitude of the steady-state current, directly indicating its K⁺-dependent nature (Fig. 1, B and D).

The basolateral membrane of principal cells is known to contain K_ir4.1/5.1 potassium channels and electrogenic Na⁺-K⁺-ATPase (40). Indeed, Fig. 2 demonstrates the immunolocalization of K_ir4.1 and K_ir5.1 channels on the basolateral membrane in mouse kidney cortex. Double immunofluorescence staining studies using antibodies against K_ir5.1 and the collecting duct-specific marker, water channel AQP2, revealed colocalization within the CCD. In addition, there is a clear staining in some tubules not positive to AQP2 (Fig. 2A). Double immunofluorescence staining with the specific marker of the distal convoluted tubule (DCT), thiazide-sensitive NCC, and K_ir4.1 revealed colocalization of K_ir4.1 with NCC in the DCT. Similarly, there is staining in tubules not positive for NCC, which most likely represents expression in the CCD (Fig. 2B). No positive staining was observed in glomeruli and proximal tubules. Therefore, both K_ir4.1 and K_ir5.1 proteins appeared localized in the basolateral membranes of CCD and DCT, as it was reported previously (20).

Both electrical currents across the apical and basolateral membranes can contribute to the whole cell current. We next determined their relative contribution upon formation of a patch on the basolateral membrane. Amiloride is known to inhibit ENaCs, thereby eliminating the driving force for the apically localized ROMK channels. However, treatment with amiloride (10 μM for 3 min) has no apparent effect on the amplitude of the whole cell current (Fig. 3, A and B), suggesting a minor contribution of the apical membrane, and the recorded whole cell current represents chiefly basolateral conductance. Consistently, inhibition of the apically localized ROMK channels with a selective antagonist, tertiapin-Q (TPNQ; 100 nM for 3 min), also does not inhibit the whole cell current (Fig. 3, A and C). These results support the concept that the basolateral membrane has much higher conductance than the apical and therefore the monitored whole cell current flows predominantly across the basolateral membrane.

To further explore the molecular identity underlying basolateral K⁺ conductance, we treated freshly isolated CCDs with the pump blocker ouabain (100 μM for 3 min). We detected just a minor trend toward a reduction of the whole cell current and depolarization, which was not statistically significant (Fig. 3, A and D). Prolonged treatment with ouabain has detrimental consequences for cellular health, which precluded patch-clamp analysis of this condition. This is consistent with the view that the activity of K_ir4.1 and K_ir5.1 potassium channels on the basolateral membrane is the chief contributor to the whole cell current in PC of freshly isolated CCDs.

There are no specific commercially available inhibitors of K_ir4.1/5.1 channels, although it was previously reported that fluoxetine (serotonin reuptake blocker) and nortriptyline (tricyclic antidepressant) are potent inhibitors of homomeric K_ir4.1 channels heterologously expressed in HEK293T cells (25, 36). Whereas fluoxetine (100 μM for 4 min) had only a minor effect (Fig. 4A), nortriptyline (100 μM for 4 min) virtually abolished macroscopic whole cell current in CCD principal cells (Fig. 4B). These results are consistent with the concept that fluoxetine affects only monomeric K_ir4.1 and nortriptyline can also target K_ir4.1/5.1 channels. Furthermore, our observations strongly argue that activity of the heteromeric K_ir4.1/5.1 is the major contributor to the K⁺-selective whole cell current recorded from the basolateral side in CCD principal cells.

Insulin acutely increases activity of basolateral K_ir4.1/5.1 in CCD principal cells in a PI3-kinase-dependent manner.

Insulin, by acting on its basolateral receptors, is known to stimulate ENaC activity in principal cells to promote Na⁺ reabsorption (35). We next asked whether insulin also modulates activity of basolateral K⁺ channels to affect transepithelial voltage. Acute insulin application significantly increases the amplitude of whole cell current from 688 ± 119 to 1,096 ± 52 pA during the steady state (Fig. 5, A and B). Insulin washout returned the current to control values of 729 ± 77 pA, indicating a reversible nature of this stimulation. Figure 5C shows the average I–V relationships of the whole cell current of PCs in control and during insulin treatment. Of note, insulin also causes a shift of the reversal potential toward more negative values (Fig. 5C), which is consistent with hyperpolarization of the basolateral membrane. Importantly, stimulatory actions of insulin were precluded in the presence of nortriptyline (Fig. 6, A and B). This suggests that activation of K_ir4.1/5.1 plays a central role in augmentations of whole cell current in response to insulin.

We next directly monitored insulin actions on this channel using a cell-attached configuration in freshly isolated CCDs. Similarly as we and other described previously (20, 45), heteromeric K_ir4.1/5.1 channels were identified by their fast kinetics, estimated single-channel conductance of ∼40 pS, and high selectivity for potassium. As demonstrated in a representative continuous experiment (Fig. 7A) and the summary graph (Fig. 7B), insulin (100 nM) reversibly increases channel P_o from 0.34 ± 0.03 to 0.45 ± 0.03. Of note, we have also detected a small insulin-induced increase (∼10%) in the unitary current amplitude (Fig. 7A, regions 1 and 2), indicating hyperpolarization of the basolateral membrane. Pretreatment with the PI3-kinase inhibitor LY294002 (20 μM for 3 min) has no effect on basal K_ir4.1/5.1 activity but abolishes the stimulatory effect of insulin (Fig. 7C), suggesting that insulin increases K_ir4.1/5.1 activity in a PI3-kinase dependent manner.

IGF-1 is less potent to stimulate K_ir4.1/5.1 activity in CCD cells.

In addition to insulin, IGF-1 also has a stimulatory effect on ENaC-mediated Na⁺ reabsorption in the CCD (35). However, application of 100 nM IGF has no effect on the whole cell current and K_ir4.1/5.1 activity (data not shown). In contrast, a higher concentration of IGF-1 (500 nM) is capable of reversibly stimulating the amplitude of whole cell current from 694 ± 79 to 1,110 ± 80 pA during the steady state (Fig. 8, A and B). The respective I–V relationships for whole cell current in control and during IGF-1 treatment are shown in Fig. 8C. A consistent stimulatory effect of a high IGF-1 concentration (500 nM) is also apparent at the single-channel level, increasing K_ir4.1/5.1 P_o from 0.31 ± 0.03 to 0.44 ± 0.03 (Fig. 9, A and B). Again, pretreatment with PI3-kinase inhibitor LY294002 (20 μM for 3 min) abolishes the effect of IGF-1 (Fig. 9C). Overall, we concluded that IGF-1, while at a higher concentration, produces a similar stimulatory action on K_ir4.1/5.1-mediated current in the CCD cells as insulin.

Insulin and IGF-1 hyperpolarize the basolateral membrane of PC in the CCD.

Our results (Figs. 5C and 8C) demonstrate that insulin and IGF-1 elicit a leftward shift of the whole cell current reversal potential. Furthermore, we detected a small increase in the K_ir4.1/5.1 unitary current amplitude during treatment with both hormones (Figs. 7A and 9A). This is consistent with an effect of insulin and IGF-1 on the basolateral membrane voltage in CCD principal cells. To directly probe this, we continuously monitored real-time change in the resting membrane potential using patch-clamp electrophysiology in current-clamp mode. Insulin (Fig. 10A) and IGF-1 (Fig. 10B) elicit an acute hyperpolarization of the basolateral membrane from −67.6 ± 1.7 to −76.0 ± 2.2 and −73.8 ± 2.2 mV, respectively (Fig. 10C). Short applications of 50 mM KCl at the end of experiments were used to test viability of cells and to verify that K⁺ conductance is the major contributor of the resting membrane potential.