Heme oxygenase metabolites inhibit tubuloglomerular feedback (TGF)

Abstract

Tubuloglomerular feedback (TGF) is the mechanism by which the macula densa (MD) senses increases in luminal NaCl concentration and sends a signal to constrict the afferent arteriole (Af-Art). The kidney expresses constitutively heme oxygenase-2 (HO-2) and low levels of HO-1. HOs release carbon monoxide (CO), biliverdin, and free iron. We hypothesized that renal HOs inhibit TGF via release of CO and biliverdin. Rabbit Af-Arts and attached MD were simultaneously microperfused in vitro. The TGF response was determined by measuring Af-Art diameter before and after increasing NaCl in the MD perfusate. When HO activity was inhibited by adding stannous mesoporphyrin (SnMP) to the MD perfusate, the TGF response increased from 2.1 ± 0.2 to 4.1 ± 0.4 μm (P = 0.003, control vs. SnMP, n = 7). When a CO-releasing molecule, (CORM-3; 50 μM), was added to the MD perfusate, the TGF response decreased by 41%, from 3.6 ± 0.3 to 2.1 ± 0.2 μm (P < 0.001, control vs. CORM-3, n = 12). When CORM-3 at 100 μM was added to the perfusate, it completely blocked the TGF response, from 4.2 ± 0.4 to −0.2 ± 0.3 μm (P < 0.001, control vs. CORM-3, n = 6). When biliverdin was added to the perfusate, the TGF response decreased by 79%, from 3.4 ± 0.3 to 0.7 ± 0.4 μm (P = 0.001, control vs. biliverdin, n = 6). The effects of SnMP and CORM-3 were not blocked by inhibition of nitric oxide synthase. We concluded that renal HO inhibits TGF probably via release of CO and biliverdin. HO regulation of TGF is a novel mechanism that could lead to a better understanding of the control of renal microcirculation and function.

tubuloglomerular feedback (TGF) is the mechanism by which the macula densa (MD) senses increases in luminal NaCl concentration and sends a signal to constrict the afferent arteriole (Af-Art) (48). Understanding the mechanism of TGF regulation is essential to a better understanding of the regulation of renal vascular resistance, glomerular filtration rate (GFR), and renal function.

Heme oxygenases (HOs) catalyze the conversion of heme to carbon monoxide (CO), biliverdin, and ferrous iron. Heme is a prosthetic group present in heme proteins such as hemoglobin and many enzymes. Two HO isoforms, HO-1 and HO-2, have been described (1). In the kidney, HO-2 is expressed constitutively mainly within the epithelial cells of the thick ascending limb, distal convoluted and connecting tubules, and the arteriolar structure (11, 16). Evidence suggests that under basal conditions CO and biliverdin generated by HO-2 are a physiological signaling molecule (12, 25). On the other hand, the basal levels of expression of HO-1 are relatively low, and its contribution to HO activity becomes apparent only under pathological conditions that cause its induction (11). HO-1 is induced by various stimuli, including angiotensin II, nitric oxide synthase (NOS) inhibition, increased reactive oxygen species, and inflammation (22, 47, 53). The HO-1 system is thought to be an antioxidant defense mechanism due to its marked upregulation during stress and its protective effect when overexpressed (34, 38, 39, 49, 52, 53). Also, there is evidence that HO products contribute to blood pressure regulation (58). Prolonged HO-1 induction antagonizes the pressor effect of the renin-angiotensin system in renovascular hypertension, probably by increasing cGMP and causing vasodilatation, decreasing oxidative stress, and by its anti-inflammatory actions (7).

The kidney is likely to be a key site of action for HO products. Metalloporphyrins, which inhibit HOs, decrease renal medullary blood flow in normal rats (59) and total renal blood flow in chronically hypoxic rats (35), suggesting that one or more HO products contribute to renal vasodilator mechanisms. In vivo studies suggest that endogenous HO products cause renal vasodilation, diuresis, and natriuresis, while inhibition of HOs decreases renal blood flow and sodium excretion (3, 46). On the other hand, HO-derived biliverdin and bilirubin are potent antioxidants (5, 50) that can directly react with superoxide (45). We have shown that superoxide in the MD potentiates the TGF response by scavenging nitric oxide (NO) (41). However, to our knowledge, there is no information addressing the role of the HO/CO/biliverdin system (HO system) in TGF regulation. We hypothesized that HOs in the nephron produce CO and biliverdin, which, in an autocrine or paracrine fashion, inhibit the TGF response by acting in the MD and/or in the Af-Art. To address this hypothesis, in vitro perfusion of a microdissected Af-Art and adherent tubular segment containing the MD was used to assess TGF responses to SnMP, an inhibitor of HOs, CO, and biliverdin, added to the MD perfusate.

MATERIALS AND METHODS

Af-Arts and attached MDs were isolated and microperfused as described previously (18, 18, 42). Young male New Zealand White rabbits (1.5–2.0 kg) were housed for 3–7 days before the beginning of the study. All procedures were approved and followed the guidelines of the Henry Ford Hospital Institutional Animal Care and Use Committee. Rabbits were anesthetized with ketamine (100 mg im), xylazine (20 mg im), and pentobarbital sodium (40 mg iv) and injected with heparin (1,000 U iv). The kidney was removed and sliced along the corticomedullary axis, and slices were placed in ice-cold MEM (GIBCO, Grand Island, NY) containing 5% BSA (Intergen). A single superficial Af-Art and intact glomerulus from each rabbit were microdissected together with adherent tubular segments consisting of portions of the thick ascending limb, MD, and early distal tubule. Samples were transferred to a temperature-regulated chamber mounted on an inverted microscope (Olympus IMT-2) with Hoffmann modulation. Both the Af-Art and the end of either the distal tubule or thick ascending limb were cannulated with an array of glass pipettes as described previously (42). Intraluminal Af-Art pressure was measured by Landis' technique, using a fine pipette introduced into the lumen through the perfusion pipette. The Af-Art was perfused with MEM supplemented with 5% BSA and (in mM) 5 NaHCO₃, 10 NaCl, 10 HEPES, and 10 NaOH. Intraluminal pressure was maintained at 60 mmHg throughout the experiment. The MD was perfused with physiological saline consisting of (in mM) 10 HEPES, 1 CaCO₃, 4 KHCO₃, 1.2 MgSO₄, 0.5 K₂HPO₄, 5.5 glucose, 0.5 Na lactate, 0.5 Na acetate, and either 10 NaCl (low NaCl) or 80 NaCl (high NaCl). Solutions are gassed with room air, which does not change the pH. The pH of each solution was between 7.3 and 7.4. The bath was similar to the arteriolar perfusate except that it contained 0.15% BSA and was exchanged continuously at a rate of 1 ml/min. Microdissection and cannulation were completed within 90 min at 8°C, after which the bath was gradually warmed to 37°C for the rest of the experiment. Once the temperature was stable, a 30-min equilibration period was allowed before any measurements were taken. Images were displayed at magnifications up to ×1,980 and recorded with a video system. Af-Art diameter was measured at 5-s intervals. Three individual measurements were taken plus or minus 5 μm around the site of maximum change in diameter using Metavue software (Universal Imaging, Carpentersville, IL).

Stannous mesoporphyrin (SnMP), an HO inhibitor, was purchased from Frontier Scientific (Logan, UT). A stock solution of 2 mM SnMP was prepared by sonication in alkaline vehicle (50 mM Na₂CO₃), resulting in a dark purple solution. Dilution of SnMP (75 μl in 10 ml) to its final concentration (15 μM) in the perfusion solution resulted in a clear, slightly purple solution with no precipitation. The perfusion solution was filtered immediately before use with a 0.45-μm syringe filter. Addition of SnMP did not significantly change the pH of the MD perfusion solution (7.3 vs. 7.33). Sodium concentration in the perfusion solutions was increased by 0.75 mM due to the sodium content in the SnMP stock solution. This increase in sodium concentration does not affect the TGF response (40). Tricarbonylchloro(glycinato)ruthenium(II), known as CO-releasing molecule 3 (CORM-3, Synthesized and provided by J. R. Falck, Dallas, TX), was used as a CO donor. This was freshly prepared before the experiments by dissolving the compound in distilled water. Inactive CORM-3 (iCORM-3) was produced by leaving CORM-3 in Krebs buffer overnight at room temperature. The resulting compound does not release CO and therefore was used as a negative control. Biliverdin was obtained from MP Biomedicals (Solon, OH). N^ω-nitro-l-arginine methyl ester (l-NAME; Sigma) at 100 μM, a NOS inhibitor, and 7-nitroindazole (7-NI; Cayman, Ann Arbor, MI) at 10 μM, an inhibitor of NOS 1, were used to inhibit NOS 1 in the MD. 7-NI was dissolved in 98% alcohol by sonication. The final alcohol concentration was 0.018%, which does not affect the TGF response (44).

Experimental protocols.

Protocols consisted of two to four consecutive TGF responses induced by high NaCl. We induced the TGF response with 80 mM NaCl, except when the treatment was expected to potentiate the response. In those protocols, we used 30–40 mM NaCl to preclude a maximum TGF response. The first TGF response was used as a control, while the subsequent TGF responses were induced in the presence of various pharmacological probes administered in the MD perfusate.

Statistics.

The TGF response was defined as the decrease in Af-Art diameter (in μm) when the NaCl concentration was switched from low to high in the MD perfusate. We compared control vs. experimental TGF responses. Since these TGF responses are done sequentially in the same preparation, first without treatment (control), then with treatment (experimental), a paired t-test was performed. In those cases where multiple comparisons were examined, Hochberg's method was used to determine significance. Values are expressed as means ± SE, and P < 0.05 was considered significant.

RESULTS

We first determined whether endogenous HOs play a role in TGF regulation in vitro. We obtained a control TGF response, added SnMP to the MD perfusate to a final concentration of 15 μM, and tested the TGF response again. The TGF response was induced with 40 instead of 80 mM NaCl to preclude a maximum TGF response, thus allowing TGF potentiation. SnMP increased the TGF response by 98%, from 2.1 ± 0.2 to 4.1 ± 0.4 μm (P = 0.003, control vs. SnMP, n = 7; Fig. 1). This shows that the HO system plays an important role in TGF regulation.

Fig. 1.

Left: effect of the heme oxygenase (HO) inhibitor stannous mesoporphyrin (SnMP; 15 μM) on the tubuloglomerular feedback (TGF) response induced by 40 mM NaCl. Right: the TGF response measured as changes in afferent arteriole (Af-Art) diameter in the absence (control) and presence of SnMP. SnMP potentiated the TGF response (n = 7).

Because HOs release CO and biliverdin from heme, we examined the effect of CO and biliverdin on the TGF response. CORM-3, a CO-releasing molecule, was added to the MD perfusate. The TGF response was induced by increasing NaCl from 10 to 80 mM. CORM-3 (50 μM) decreased the TGF response by 41%, from 3.6 ± 0.3 to 2.1 ± 0.2 μm (P < 0.001, control vs. CORM-3, n = 12; Fig. 2). When CORM-3 was added at higher concentration (100 μM), it completely blocked the TGF response (100%), from 4.2 ± 0.4 to −0.2 ± 0.3 μm (P < 0.001, control vs. CORM-3, n = 6; Fig. 3). To rule out the possibility that CO released by high concentrations of CORM-3 may have injured the MD cells, we tested whether the effect of CORM-3 was reversible. After CORM-3 was removed while MD perfusion was continued with 80 mM NaCl, the TGF response was restored to 4.7 ± 1.3 μm, n = 3 (Fig. 3). These data suggest that inhibition of the TGF response by CO did not damage the MD cells.

Fig. 2.

Left: effect of CO-releasing molecule 3 (CORM-3; 50 μM) on the TGF response induced by 80 mM NaCl. Right: the TGF response measured as changes in Af-Art diameter in the absence (control) and presence of CORM-3. CORM-3 (50 μM) partially inhibited the TGF response (n = 12).

Fig. 3.

Left: effect of CORM-3 (100 μM) on the TGF response induced by 80 mM NaCl. Right: the TGF response measured as changes in Af-Art diameter in the absence (control) and presence of CORM-3. CORM-3 (100 μM) completely inhibited the TGF response (n = 6). The TGF response was completely restored after removal of CORM-3 (n = 3).

In a separate experiment, inactive CORM-3 was used as a negative control. The TGF response was induced by 80 mM NaCl. During the control TGF response, Af-Art diameter decreased by 3.7 ± 0.4 μm. After the addition of inactive CORM-3 (100 μM), Af-Art diameter decreased by 3.6 ± 0.3 μm; n = 4 (Fig. 4). These data indicate that inactive CORM-3 did not alter the TGF response.

Fig. 4.

Left: effect of inactive CORM-3 (iCORM; 100 μM) on the TGF response induced by 80 mM NaCl. Right: the TGF response measured as changes in Af-Art diameter in the absence (control) and presence of iCORM-3. iCORM-3 did not alter the TGF response (n = 4).

We next tested whether biliverdin participates in TGF regulation. Biliverdin (5 μM) was added to the MD perfusate, and the TGF response was induced by increasing NaCl from 10 to 80 mM. Biliverdin decreased the TGF response by 79%, from 3.4 ± 0.3 to 0.7 ± 0.4 μm (P = 0.001, control vs. biliverdin, n = 6; Fig. 5). After biliverdin was removed while MD perfusion was continued with 80 mM NaCl, the TGF response was restored to 2.7 ± 1.6 μm, n = 3 (Fig. 5). These data suggest that biliverdin in the MD inhibits the TGF response and that this inhibition is reversible.

Fig. 5.

Left: effect of biliverdin (5 μM) on the TGF response induced by 80 mM NaCl. Right: the TGF response measured as changes in Af-Art diameter in the absence (control) and presence of biliverdin. Biliverdin inhibited the TGF response (n = 6). The TGF response was completely restored after removal of biliverdin (n = 3).

Since it has been reported that metalloporphyrins at high concentrations may inhibit NOS, and since NOS inhibition in the MD potentiates the TGF response (4, 43, 44, 56), we tested whether TGF potentiation by SnMP is due to inhibition of NOS 1 in the MD. The TGF response was induced with 30 mM NaCl instead of 80 to preclude a maximum response. This concentration of NaCl resulted in a small but significant decrease in Af-Art diameter, of 1.0 ± 0.1 (P < 0.001). As previously reported (18), the NOS inhibitor l-NAME potentiated the TGF response. l-NAME (100 μM) increased the TGF response by 156%, from 1.0 ± 0.1 to 2.6 ± 0.2 μm (P < 0.001, control vs. l-NAME). Addition of SnMP while l-NAME was continued caused the TGF response to further increase by 63%, from 2.6 ± 0.2 to 4.2 ± 0.4 μm (P = 0.01, l-NAME vs. l-NAME+SnMP, n = 8; Fig. 6). These data suggest that most of the effect of SnMP is independent of NOS activity.

Fig. 6.

Effect of inhibition of tubular nitric oxide synthase [NOS; N^ω-nitro-l-arginine methyl ester(l-NAME), 100 μM] on potentiation of the TGF response by the HO inhibitor SnMP (15 μM). Left: 3 consecutive TGF responses induced by 30 mM NaCl in the presence of vehicle (control), l-NAME, and l-NAME plus SnMP, respectively. Right: the TGF response measured as changes in Af-Art diameter in control and in the presence of l-NAME and l-NAME plus SnMP. l-NAME did not prevent the potentiation of TGF by SnMP (n = 8).

Because it has been reported that part of the vasodilatory effect of CO may be due to release of NO (51), we tested whether the inhibition of the TGF response by CORM-3 is due to increased NO release. We induced four consecutive TGF responses by increasing NaCl from 10 to 80 mM in the presence of vehicle, CORM-3 (100 μM), the NOS 1 inhibitor 7-NI (10 μM), or CORM-3+7-NI. During the control period, the TGF response was 2.7 ± 0.1 μm. In the presence of CORM-3, the TGF response was 0.3 ± 0.1 μm (P < 0.001, control vs. CORM-3). In the presence of 7-NI, the TGF response was 4.6 ± 0.5 μm (P = 0.02, control vs. 7-NI). Finally, in the presence of CORM-3+7-NI, the TGF response was 0.01 ± 0.3 μm (P = 0.003, 7-NI vs. 7-NI+CORM-3, n = 5; Fig. 7). The TGF response was completely blocked by CORM-3, despite the presence of 7-NI, suggesting that its effect is independent of the generation of NO.

Fig. 7.

Effect of inhibition of tubular NOS [7-nitroindazole (7-NI), 10 μM] on inhibition of the TGF response by the CO-releasing molecule CORM-3 (100 μM). Left: 4 consecutive TGF responses induced by 80 mM NaCl in the presence of vehicle (control), CORM-3, 7-NI, and 7-NI plus CORM-3, respectively. Right: the TGF response measured as changes in Af-Art diameter in control and in the presence of CORM-3, 7-NI, and 7-NI plus CORM-3. 7-NI did not prevent inhibition of TGF by CORM-3 (n = 5).

DISCUSSION

Using isolated perfused MDs and Af-Arts, we have obtained evidence supporting our hypothesis that the HO system inhibits the TGF response, because inhibiting tubular HOs with SnMP potentiated the TGF response. Furthermore, adding exogenous CO or biliverdin to the MD perfusate inhibited the TGF response. Thus, taken together, these data suggest that the HO/CO/biliverdin system plays a significant role in TGF regulation.

Expression of HOs in the MD has not been studied; it could be that a HO isoform is expressed in the MD, probably HO-2, which might regulate TGF in an autocrine manner by releasing CO and/or biliverdin. However, it is also possible that products of HO produced in other segments of the nephron acting in a paracrine manner regulate TGF. HO-2 is constitutively expressed in the vascular and tubular structures of the kidney (11, 15, 46, 47, 59). In the nephron, HO-2 mRNA and protein are localized in the epithelial cells of the thick ascending limb, distal convoluted tubule, connecting tubule cells, and principal cells of the collecting duct (11, 16). Evidence suggests that, under basal conditions, CO and biliverdin generated by HO-2 is a physiological signaling molecule (12, 25). We have previously shown that another gas, NO, produced by NOS in the thick ascending limb, acts as paracrine factor reaching the MD and inhibiting TGF (54). Thus it is possible that HO-2 on one of these renal structures produces CO and biliverdin that, acting as paracrine factors, participate in the regulation of TGF. On the other hand, the basal levels of expression of HO-1 are relatively low, and its contribution to HO activity becomes apparent only under pathological conditions that cause its induction (11). HO-1 is induced by various stimuli, including angiotensin II, NOS inhibition, increased reactive oxygen species, and inflammation (22, 47, 53).

In the present work, we used a specific inhibitor of HOs to study the role of endogenous CO and biliverdin in TGF regulation. Specific HO inhibitors are important tools in the assessment of the role of CO and biliverdin as messenger molecules. Metalloporphyrins have been shown to inhibit HO, and their potency is affected by the metal cation associated with the porphyrin ring as well by different ring substituents. We chose SnMP since it is a selective inhibitor of HO activity and does not inhibit NOS activity (29). Tin incorporated in the porphyrin ring structure to form SnMP is extremely stable, and there is no known physiological mechanism by which the element can be removed from the porphyrin complex (24). Our study using SnMP provides direct evidence that a functional HO is expressed in the nephron and regulates TGF, since HO inhibition in the tubular compartment potentiates the TGF response. Interestingly, HO inhibition in the nephron does not affect Af-Art diameter when the MD is perfused with 10 mM NaCl. The mechanism of HO regulation in the nephron is unknown. It could be that HOs are activated during TGF, since TGF increases intracellular pH (6) and PKC activity, two known HO activators (12, 26). Similarly, we have previously shown that increased intracellular pH at the MD activates NOS 1 during TGF (23).

Since HOs generate CO, biliverdin, and Fe(II), we tested the effect of exogenous CO and biliverdin on the TGF response. CO is a versatile signaling molecule with essential regulatory roles in a variety of physiological and pathophysiological processes (27, 32, 37). In our study, CO showed a dose-dependent inhibition of the TGF response, with high doses inhibiting the TGF response completely. This effect was not due to injury of the juxtaglomerular apparatus, since on withdrawal of CORM-3 from the perfusate, the TGF response was immediately restored. CORM-3 is a CO-releasing molecule that liberates CO in a concentration-dependent manner in vitro and in vivo (31, 33). It is possible that inhibition of the TGF response by exogenous CORM-3 and biliverdin is a pharmacological effect. However, since inhibition of HO potentiates the TGF response, probably by decreasing endogenous CO and/or biliverdin, these HO metabolites may play a physiological role in inhibiting the TGF response. It could be that CORM-3 acts independently of CO release. However, inactive CORM-3, the resulting molecule after release of CO from CORM-3 (14), had no effect on the TGF response.

In addition to CO, HOs also release biliverdin, which is rapidly metabolized to bilirubin by biliverdin reductase (2, 57). Our data indicate that biliverdin inhibits the TGF response induced by high NaCl. Thus the effects of HO inhibition may be attributed, at least in part, to reduced biliverdin release. It is not clear how biliverdin blocks the TGF response. Biliverdin and bilirubin are potent antioxidants (5, 50), and there is evidence that they may decrease Na reabsorption (9, 10, 13, 28). We have shown that superoxide in the MD potentiates the TGF response by scavenging NO (41). Therefore, it is possible that biliverdin in the MD inhibits the TGF response by decreasing superoxide, thereby increasing bioavailable NO and reducing type 2 Na-K-2Cl cotransporter-dependent sodium transport (36).

It could be speculated that administration of SnMP, CORM-3, or biliverdin in the MD perfusate may diffuse through the bath and act directly in the Af-Art, causing vasoconstriction or vasodilation. This is highly unlikely, since the MD perfusion rate is ∼20 nl/min, while the exchange rate of bath perfusate is 1 ml/min. Effluents from the MD (SnMP, CO, or biliverdin) are immediately diluted 50,000 times and washed out before they can act on other structures (17). In addition, it has been reported that direct application of a HO inhibitor on the Af-Art does not cause constriction (8).

Since it has been reported that metalloporphyrins at high concentrations may inhibit NOS, and since NOS inhibition in the MD potentiates the TGF response (4, 43, 44, 56), we studied the role of NOS in the potentiation of the TGF response by SnMP. In the presence of l-NAME, the potentiation of the TGF response by SnMP was still present, indicating that at least part, if not all, of the effect of SnMP is independent of NOS 1. Furthermore, as mentioned, SnMP is a selective inhibitor of HOs and does not inhibit NOS activity at the doses we used (29).

Thorup et al. (51) reported that low concentrations of CO applied directly in the Af-Art induce significant dilatation and that part of this effect may be due to a release of NO. Thus we studied whether part of the inhibition of the TGF response by CO is due to the release of NO. We found that in the presence of a NOS 1 inhibitor, 7-NI, TGF response inhibition by CORM-3 was not altered.

There is evidence that the renal HO system may have a natriuretic effect, since infusion of exogenous heme (the substrate for HOs) induces natriuresis and these effects are HO dependent, because they are blocked in the presence of an HO inhibitor (46). In vivo the HO system shifts the pressure-natriuresis curve to the left (21, 59), possibly by inhibiting tubular Na reabsorption and/or the TGF response. It could be that TGF regulation by the HO system is due to inhibition of Na transport by the type 2 Na-K-2Cl cotransporter in the MD. This acts as a chemical detector, sensing increased luminal NaCl concentrations (48) and initiating TGF signaling.

There is also evidence that CO may play a role in the regulation of the renal microcirculation. CO is a vasodilator (20, 58), an effect linked to activation of soluble guanylyl cyclase (30) and/or stimulation of Ca-activated potassium channels (55). Endogenous CO production attenuates the effect of vasoconstrictors (19) in the renal interlobular arteries. Exogenous CO, when applied directly to the Af-Art, induces vasodilation (8, 51).

We conclude that renal HOs regulate (inhibit) the TGF response via release of CO and biliverdin. The effects of HO inhibition and CO are independent of NOS. HO regulation of the TGF response is a novel mechanism that could lead to a better understanding of the control of renal microcirculation and function in physiological and pathophysiological conditions.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grant HL-28982.