NADPH oxidase 4 mediates flow-induced superoxide production in thick ascending limbs

Abstract

We previously showed that luminal flow stimulates thick ascending limb (TAL) superoxide (O₂⁻) production by stretching epithelial cells and increasing NaCl transport, and reported that the major source of flow-induced O₂⁻ is NADPH oxidase (Nox). However, the specific Nox isoform involved is unknown. Of the three isoforms expressed in the kidney—Nox1, Nox2, and Nox4—we hypothesized that Nox4 is responsible for flow-induced O₂⁻ production in TALs. Measurable flow-induced O₂⁻ production at physiological flow rates of 0, 5, 10, and 20 nl/min was 5 ± 1, 9 ± 2, 36 ± 6, and 66 ± 8 AU/s, respectively. RT-PCR detected mRNA for all three Nox isoforms in the TAL. The order of RNA abundance was Nox2 > Nox4 >>> Nox1. Since all three isoforms are expressed in TALs and pharmacological inhibitors are not selective, we used rats transduced with siRNA and knockout mice. Nox4 siRNA knocked down Nox4 mRNA expression by 63 ± 7% but did not reduce Nox1 or Nox2 mRNA. Flow-induced O₂⁻ was 18 ± 9 AU/s in TALs transduced with Nox4 siRNA compared with 77 ± 9 AU/s in tubules transduced with scrambled siRNA. Flow-induced O₂⁻ was 81 ± 5 AU/s in Nox2 knockout mice compared with 83 ± 13 AU/s in wild-type mice. In TALs transduced with Nox1 siRNA, flow-induced O₂⁻ was 82 ± 7 AU/s. We conclude that Nox4 mediates flow-induced O₂⁻ production in TALs.

superoxide (O₂⁻) regulates renal function (51, 58). This reactive oxygen species (ROS) favors salt and water retention and has been implicated in salt-sensitive hypertension (23, 30, 49). The thick ascending limb of the loop of Henle (TAL) is the main source of medullary O₂⁻ (28, 29, 58), and O₂⁻ stimulates NaCl reabsorption in this segment (21, 34, 48). The TAL regulates urinary volume and sodium excretion. Therefore, factors that affect TAL function may lead to inappropriate NaCl retention and result in salt-sensitive hypertension (20, 22, 41).

Flow of urine as it forms along the nephron is highly variable (8, 43). Luminal flow through the TAL normally varies from 0 to 20 nl/min (8, 42, 43). Acute variations in luminal flow are a result of tubuloglomerular feedback (27) and peristalsis of the renal pelvis (8, 43). Luminal flow also varies chronically due to volume changes (1, 42), a high-salt diet (32), hypertension (2, 4, 6, 40, 56), and diabetes (38, 55).

We found that luminal flow stimulates O₂⁻ production in TALs by enhancing Na-K-2Cl cotransporter activity (18) and stretching epithelial cells (12), and we determined that the source of O₂⁻ is NADPH oxidase (Nox) (19). The Nox family of superoxide-generating enzymes consists of at least seven isoforms of the major catalytic subunit (24). The adult kidney expresses Nox1, Nox2, and Nox4 (14). Nox2 mRNA (14, 28) and protein (14, 54) and Nox4 protein (54) have been detected in TALs, whereas thus far Nox1 has only been found in the proximal tubules of the kidney cortex (14, 44). Each of the three isoforms has been implicated in signaling pathways involving O₂⁻ stimulation in the kidney (14). However, it is unclear which of these isoforms is involved in flow-stimulated O₂⁻.

Nox4 was originally named Renox because of its high expression in the kidney. We hypothesized that Nox4 is the source of O₂⁻ production stimulated by increases in luminal flow. To test this hypothesis, we examined the Nox isoforms expressed in TALs and used knockout animal models and adenoviral delivery of Nox isoform-specific siRNAs to determine which isoforms are involved in flow-stimulated O₂⁻. Our findings indicate that 1) TALs predominantly express Nox2 and Nox4 mRNA, and to a much lower extent Nox1 mRNA; and 2) only Nox4 is involved in flow-stimulated O₂⁻ production.

MATERIALS AND METHODS

Chemicals and solutions.

Dihydroethidium was purchased from Molecular Probes (Eugene, OR), RNeasy Plus mini kits and OneStep RT-PCR kits from Qiagen (Valencia, CA), and Nox isoform-specific PCR primers from TIB Molbiol (Adelphia, NJ). Custom oligonucleotides used to construct adenoviruses carrying Nox1- and Nox4-specific siRNA were purchased from Eurofins MWG Operon (Huntsville, AL). The composition of the physiological saline used to perfuse and bathe the TALs was (in mM) 130 NaCl, 2.5 NaH₂PO₄, 4 KCl, 1.2 MgSO₄, 6 l-alanine, 1 trisodium citrate, 5.5 glucose, 2 calcium dilactate, and 10 HEPES, pH 7.4 at 37°C. All solutions were adjusted to 290 ± 3 mosmol/kgH₂O.

Animals.

Male Sprague-Dawley rats (Charles River Breeding Laboratories, Wilmington, MA) were maintained on a diet containing 0.22% sodium and 1.1% potassium (Purina, Richmond, IN) for at least 3 days. Male Nox2 knockout (−/−) and C57BL/6J wild-type mice (Jackson Laboratory, Bar Harbor, ME) were fed normal rodent chow containing 0.4% sodium. All animal protocols were approved by the Institutional Animal Care and Use Committee of Henry Ford Hospital in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

RT-PCR.

Separate TAL suspensions from the adenovirus-injected left kidney and noninjected right kidney were prepared as described previously (47). Total RNA was isolated from the suspensions using an RNeasy Plus mini kit and subsequently used for RT-PCR analysis using a OneStep RT-PCR kit according to the manufacturer's instructions. Previously published Nox isoform-specific PCR primers (50) were used to assay mRNA expression. Thermal cycler conditions for OneStep RT-PCR were 1) reverse transcription: 30 min at 50°C; 2) HotStarTaq DNA polymerase activation at 95°C for 15 min; 3) cycling at 94°C for 15 s, 60°C for 5 s, and 72°C for 1 min; and 4) final extension at 72°C for 5 min. The total number of cycles used for detection and comparison of relative abundance of Nox isoforms in rat TALs (see Fig. 2) was 35 cycles using 1 μg RNA. However, because of the differences observed in mRNA expression among the three Nox isoforms in TALs, the number of cycles and amount of template RNA were adjusted for the purpose of evaluating knockdown of each particular Nox mRNA by an adenovirally delivered Nox isoform-specific siRNA (see Fig. 3). For Nox1 mRNA knockdown analysis, the PCR primers were 5′-GGAGTCCTCATTTTGTGGGGCAACC-3′ (forward) and 5′-GGCACCCGTCTCTCTACAAATCCAGT-3′ (reverse). The total number of cycles used for Nox1, Nox2, and Nox4 was 35, 30, and 31, respectively, and the amount of RNA used was 400, 100, and 200 ng. PCR products were visualized on 1.5% agarose gels and stained with ethidium bromide (0.5 μg/ml). Band densities of the expected 560-, 605-, and 409-bp amplicons for Nox1, Nox2, and Nox4, respectively, were analyzed.

Adenovirus construction.

The target (sense) sequence of the siRNA against Nox4 was 5′-CCGTTTGCATCGATACTAA-3′ (bases 1358–1376 in Genbank Accession number NM_053524 for rat Nox4 mRNA). The target sequence of Nox1 siRNA (31) was 5′- AGATCTATTTCTACTGGAT-3′ (bases 1419–1437 in Genbank Accession number NM_053683). A 63-base annealed oligonucleotide siRNA cassette containing the siRNA sequence, a 5′ Afl II overhang, a 3′ Spe I overhang, a transcriptional start, and termination sequence, and a hairpin loop region was inserted into the adenoviral shuttle vector pVQAdMIGHTY, which contains the H1 promoter, and sent to ViraQuest (North Liberty, IA) for adenovirus production. For the negative control, a nonsilencing sequence (Qiagen) was similarly inserted into the shuttle vector for adenovirus construction.

Adenoviral gene delivery.

Rat medullary TALs (mTALs) were transduced in vivo with an adenovirus carrying a Nox4 siRNA (AdNox4siRNA), a Nox1 siRNA (AdNox1siRNA), or a nonsilencing sequence (AdscrsiRNA) as reported previously (35, 36). The left kidney of a rat weighing 85–105 g was exposed via a flank incision and the left renal artery and vein were clamped. Four 20-μl virus injections (4 × 10¹⁰ PFU/ml) at a rate of 20 μl/min were made by inserting the needle perpendicular to the renal capsule, parallel to the medullary rays, and directed toward the medulla. Injection points 2.5 mm apart were selected along the longitudinal axis of the kidney. The clamp was released and the kidney was returned to the abdominal cavity. The muscle was sutured and the skin was closed with staples. mTALs were dissected 6–7 days after injection, when maximal knockdown of Nox4 mRNA (63 ± 7%; n = 4, P < 0.003) was observed. For maximal Nox1 mRNA knockdown (59 ± 6%; n = 3, P < 0.01), a total of 100 μl AdNox1siRNA was used, and TALs were dissected 8 days after injection.

Isolation and perfusion of rat TALs.

On the day of the experiment, rats weighing 120–150 g or mice weighing 20–25 g were anesthetized with ketamine and xylazine (100 and 20 mg/kg body wt ip, respectively). After the abdominal cavity was opened, the left kidney was bathed in ice-cold saline and removed. Sagittal slices were placed in oxygenated physiological saline and mTALs were dissected from the outer medulla under a stereomicroscope at 4–10°C. Tubules measuring 0.5–1.0 mm were transferred to a temperature-regulated chamber and perfused using concentric glass pipettes at 37 ± 1°C as described previously (11). Except for the flow rates used for Fig. 1, we perfused tubules at a rate of 20 nl/min.

Fig. 1.

Superoxide (O₂⁻) production by the thick ascending limb in response to various luminal flow rates. P < 0.0003 for 10 nl/min and P < 0.0001 for 20 nl/min (compared with 0 nl/min); n = 4 to 6.

Measurement of O₂⁻ using dihydroethidium.

Isolated TALs were loaded with 10 μM dihydroethidium in physiological saline for 15 min and then washed in dye-free solution for 20 min. O₂⁻ converts dihydroethidium to oxyethidium (10, 57). Oxyethidium and dihydroethidium were excited using 490- and 365-nm light, respectively. Fluorescence emitted between 520–600 nm (oxyethidium) and 400–450 nm (dihydroethidium) was imaged digitally with an image intensifier connected to a CCD camera and recorded with a Metafluor imaging system (Universal Imaging, West Chester, PA). Fluorescence from regions of interest was measured every 5 s for 12 measurements. Regression analysis of the fluorescence ratios for each measurement period was performed and differences in slopes were evaluated. An increased rate of change in the oxyethidium/dihydroethidium fluorescence ratio is a measure of increased net O₂⁻ production (production − degradation). We will refer to O₂⁻ measurements in this study as O₂⁻ production.

Statistical analysis.

Results are expressed as means ± SE. Data were evaluated with Student's t-test, taking P < 0.05 as significant.

RESULTS

Physiological luminal flow rates stimulate O₂⁻ production.

Physiological luminal flow rates in the TAL can vary from 0 to 20 nl/min. To characterize the relationship between flow and O₂⁻ production, we subjected isolated rat TALs to luminal flow rates of 0, 5, 10, and 20 nl/min and measured O₂⁻ production. We found it was 5 ± 1 AU/s (n = 6) at 0 nl/min, 9 ± 2 AU/s at 5 nl/min (n = 6), 36 ± 6 AU/s at 10 nl/min (P < 0.0003; n = 4 vs. no flow), and 66 ± 8 AU/s at 20 nl/min (P < 0.0001; n = 4 vs. no flow). Therefore, O₂⁻ production increased linearly with flow at rates 5 nl/min and above. However, O₂⁻ production at 5 nl/min was no different from basal no flow conditions (Fig. 1). Whether 5 nl/ml represents an actual point at which there is no distinguishable increase in O₂⁻ production in response to flow or instead is due to limitations in the method of measuring O₂⁻ production is unclear.

TALs express Nox1, Nox2, and Nox4 mRNA.

To identify the Nox isoform(s) involved in flow-stimulated O₂⁻, we first studied which isoforms are expressed in rat TALs by using RT-PCR. We found mRNA of each of the three isoforms in the TAL, although Nox1 was expressed far less than Nox2 or Nox4 (Fig. 2). To demonstrate that RT-PCR could be used as a semiquantitative method, we performed a serial dilution of total RNA to generate a range of starting template from 31 to 250 ng. We found that Nox4 mRNA expression was linear to the amount of RNA template (correlation coefficient = 0.9955). Additionally, because Nox1 was barely detectable in our TAL preparation, we confirmed that the Nox1 PCR primers were working properly by using them with total RNA from a VSMC cell line, in which Nox1 is highly expressed. Therefore, we can conclude that the relative abundance of the three isoforms is Nox2 > Nox4 >>> Nox1.

Fig. 2.

RT-PCR analysis of NADPH oxidase (Nox) isoforms expressed in the thick ascending limb. MW = 100-bp DNA molecular weight ladder.

In vivo transduction of TALs with adenovirus carrying Nox4 siRNA specifically and efficiently knocks down Nox4.

Of the three Nox isoforms expressed in TALs, Nox4 is the most highly expressed in the kidney. Therefore, we first chose to study the role of Nox4 in flow-stimulated O₂⁻ production. Because there are no Nox4 −/− animal models, we transduced the outer medulla of the rat kidney in vivo with an adenovirus carrying a Nox4 siRNA (AdNox4siRNA). To establish the validity of our siRNA sequence, we transduced TALs with AdNox4siRNA and measured the effect on Nox1, Nox2, and Nox4 mRNA expression (Fig. 3). We found that the Nox4 siRNA knocked down Nox4 mRNA in the TALs from the virus-injected left kidney by 63 ± 7% (P < 0.003) compared with the noninjected right kidney, whereas it had no effect on Nox1 (16 ± 7%) or Nox2 mRNA (44 ± 39%). Therefore, the efficacy and specificity of the Nox4 siRNA were suitable to study the role of Nox4 in flow-induced O₂⁻ production.

Fig. 3.

Adenovirus-delivered Nox4 siRNA specificity in the thick ascending limb.

In vivo knockdown of Nox4 inhibits flow-induced O₂⁻.

Next, we examined the effect of knockdown of Nox4 on flow-stimulated O₂⁻. In isolated TALs transduced with a control adenovirus that contained a scrambled nonsilencing sequence (AdscrsiRNA), O₂⁻ production increased from 10 ± 3 to 87 ± 12 AU/s (Δ = 77 ± 9 AU/s; P < 0.002; n = 5) when subjected to flow. However, in TALs transduced with the AdNox4siRNA, flow did not significantly increase O₂⁻ production (7 ± 2 to 25 ± 9 AU/s; Δ = 18 ± 9 AU/s; n = 6; Fig. 4). Transduction of TALs with AdscrsiRNA did not affect Nox4 mRNA expression. Therefore, knockdown of Nox4 completely inhibited flow-stimulated O₂⁻ production.

Fig. 4.

Flow-induced O₂⁻ in Nox4 siRNA-transduced thick ascending limbs. Neg con siRNA, transduction with nonsilencing AdscrsiRNA. Nox4 siRNA, transduction with AdNox4siRNA.

Nox2 is not involved in flow-induced O₂⁻.

We next assessed the possible involvement of Nox2 by measuring flow-stimulated O₂⁻ in Nox2 −/− mice (Fig. 5). Initiation of flow increased O₂⁻ production from 15 ± 2 to 96 ± 6 AU/s (Δ = 81 ± 5 AU/s; P < 0.0006, n = 4) in Nox2 −/− mice. In wild-type mice luminal flow enhanced O₂⁻ production from 19 ± 7 to 102 ± 20 AU/s (Δ = 83 ± 13 AU/s; P < 0.03; n = 3). To determine whether there are compensatory changes in Nox4 expression in Nox2 −/− mice, we compared Nox4 mRNA in wild-type and Nox2 −/− mice. We found that there was no difference. These data indicate that although Nox2 mRNA is highly expressed in TALs, it is not involved in production of O₂⁻ in response to luminal flow.

Fig. 5.

Flow-stimulated O₂⁻ production in wild-type and Nox2 knockout (−/−) mice.

In vivo knockdown of Nox1 does not affect flow-induced O₂⁻.

In endothelial cells a “vicious cycle” in which ROS-activated Nox stimulates Nox-dependent ROS production has been proposed (7). Therefore, it is conceivable that another source of O₂⁻, such as Nox1, which is also present in TALs, could be linked to stimulation of Nox4 to cause enhanced O₂⁻ production. To make sure that the flow-induced increase in O₂⁻ derived from Nox4 is not coupled to Nox1-stimulated O₂⁻, we tested rat TALs transduced with an adenovirus carrying an siRNA against Nox1 (AdNox1siRNA). The Nox1 siRNA had no effect on flow-stimulated O₂⁻ production (Fig. 6). In control experiments (nontransduced TALs), flow increased O₂⁻ production from 7 ± 2 to 98 ± 14 AU/s (Δ = 91 ± 13; P < 0.02). In Nox1siRNA-transduced TALs, flow stimulated O₂⁻ production to a similar extent (from 5 ± 2 to 87 ± 9 AU/s; Δ = 82 ± 7; P < 0.002). Therefore, knockdown of Nox1 did not affect flow-stimulated O₂⁻ production. The Nox1 siRNA knocks down Nox1 mRNA to a similar extent (59 ± 6) as the Nox4 siRNA we used to knockdown Nox4 mRNA (63 ± 7%). Knockdown of Nox1 siRNA does not affect Nox4 or Nox1 expression.

Fig. 6.

Flow-induced O₂⁻ in Nox1 siRNA-transduced thick ascending limbs. Con, nontransduced. Nox1 siRNA, transduction with AdNox1siRNA.

DISCUSSION

We demonstrated that O₂⁻ production increases in response to various flow rates. The flow-stimulated increases in O₂⁻ at these flow rates are similar to what we observed previously (18). It appears that the break point for response to flow is approximately at 5 nl/min. O₂⁻ production at 5 nl/min was not significantly different than at 0 nl/min. Whether this reflects a true maximum flow rate at which O₂⁻ is not produced or merely a limitation of our method of measuring low O₂⁻ production rates is unclear. However, we estimate that the threshold for triggering O₂⁻ production is somewhere around 5 nl/min.

Although previous studies showed that the adult kidney expresses Nox1, Nox2, and Nox4, the isoforms present and their relative abundance in specific segments of the kidney such as the TAL are not well-defined. We were able to detect in TALs the mRNA of each of the three isoforms. This is consistent with others who reported that TALs express Nox2 (14, 28, 44, 54) and Nox4 (14, 44, 54). We also found Nox1 mRNA is present in TALs, albeit to a much lesser extent than Nox2 and Nox4. To our knowledge, we are the first to show that Nox1 mRNA is expressed in the TAL.

Relative band density can be used to deduce relative expression of each mRNA, because the same amount of RNA and number of cycles were used to detect each isoform relative to one another, and the reactions for each isoform were run together and therefore under the same conditions. Furthermore, assuming the same efficiency for each set of primers in the reverse transcription and PCR reactions and given that the PCR products are fairly similar in size, we can conclude that the relative abundance of mRNA in TALs is Nox2 > Nox4 >>> Nox1.

We found the effect of flow on O₂⁻ production in TALs transduced with Nox4 siRNA was inhibited, whereas it was not affected in Nox2 −/− mice or in TALs transduced with Nox1 siRNA. These data indicate that the increase in O₂⁻ produced in response to increased luminal flow is derived from Nox4 and not Nox2 or Nox1. One might say that identification of Nox4 as the isoform involved in flow-mediated O₂⁻ was expected because Nox4, which was originally termed Renox due to its high expression in the kidney (13), is detected in renal tubular epithelium (44). Although high expression of an isoform does not necessarily equate to physiological importance, Nox4 activity has been reported to be determined by mRNA levels (46).

The mechanism whereby Nox4 activity is regulated is unclear. The most characterized isoform of the NADPH oxidase family is the one found in phagocytic cells. It is composed of several subunits: 1) the major catalytic subunit (gp91^phox, also known as Nox2) and p22^phox, which are integral membrane proteins that form the flavocytochrome b₅₅₈; and 2) p47^phox, p67^phox, p40^phox, and the small GTPase Rac1, which reside in the cytosol until activated. In the current model of activation, a stimulus phosphorylates and activates p47^phox, which organizes the other cytosolic subunits into a complex that translocates to the membrane to allow binding to the flavocytochrome and form an active Nox (3, 25). Nox4 is reportedly constitutively active (46). However, our data suggest that flow acutely stimulates Nox4 activity. Our laboratory also found that angiotensin II rapidly increases O₂⁻ in a Nox4-dependent manner. These data are consistent with others who showed that Nox4 is acutely stimulated in proximal tubules (33), mesangial cells (16, 17), and cardiac fibroblasts (5). The mechanism by which Nox4 is upregulated in the TAL rather than other Nox isoforms is unknown. The answer may lie in the differential distribution of Nox isoforms within the cell. The Nox4/p22^phox complex has been localized to the endoplasmic reticulum in endothelial cells and epithelial cell lines, whereas the flavocytochrome complex of Nox2 and Nox1 are typically found in the plasma membrane (3). Another possibility is a scheme that involves regulation of mechanosensors, because flow-induced increases in cellular stretch regulate Nox activity in many cell types, including TALs (12) and vascular cells (52).

In endothelial cells, ROS-induced ROS production occurs in which ROS derived from one source stimulates ROS production by a different source to create a “vicious cycle” (7). O₂⁻ production in TALs may also represent such a feed-forward system. We showed that flow stimulates O₂⁻ production via PKC-α-mediated activation of Nox (19). O₂⁻ also activates PKC-α (48), which can further stimulate O₂⁻ production. Therefore, it is conceivable that a signaling pathway exists in TALs in which Nox1 or Nox2 is activated by flow to produce O₂⁻, which in turn could stimulate Nox4 to produce even more O₂⁻. However, global knockout of Nox2 in the Nox2 −/− mice had no effect on flow-induced O₂⁻, thus eliminating any possible involvement of Nox2. Transduction of TALs with Nox1 siRNA also had no effect on flow-induced O₂⁻. Therefore, we can rule out the possibility that Nox1 may be involved in flow-stimulated O₂⁻. Additionally, given that Nox1 mRNA is apparently only weakly expressed in TALs and we observed complete inhibition of flow-stimulated O₂⁻ in TALs transduced with the Nox4 siRNA, it is unlikely that Nox1 plays any role in flow-induced O₂⁻.

The increase in O₂⁻ derived from Nox and its implication in the pathophysiology of diseases such as hypertension and diabetes have been studied extensively in the vasculature (26, 37, 39, 51). The role of Nox4 in the oxidative stress and renal damage associated with diabetes have also been reported (15, 45). However, the effect of O₂⁻ in the TAL has not been as widely studied. One group showed that in type 1 diabetes PKC-dependent Nox activation increases O₂⁻ production and enhances Na transport (53–55). These reports are consistent with our data that implicate the importance of Nox-derived O₂⁻ in conditions in which luminal flow is acutely or chronically enhanced, such as high-salt intake (9, 32), volume expansion (1, 42), and the osmotic diuresis associated with hypertension (2, 4, 6, 40, 56) and diabetes (38, 55).

In summary, we found that luminal flow stimulates Nox4-dependent O₂⁻ production by the TAL. The increased Na⁺ reabsorption caused by enhanced O₂⁻ in the TAL may be important in the pathogenesis of hypertension, diabetes, and other diseases associated with Na⁺ retention and chronically increased flow. Pharmacological agents that target Nox in the TAL may lead to new therapeutic approaches to treating disorders associated with oxidative stress.

GRANTS

This work was supported by National Institutes of Health Grants HL28982, HL70985, and HL90550 to J. L. Garvin.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: N.J.H. performed experiments; N.J.H. analyzed data; N.J.H. and J.L.G. interpreted results of experiments; N.J.H. prepared figures; N.J.H. drafted manuscript; N.J.H. and J.L.G. edited and revised manuscript; N.J.H. and J.L.G. approved final version of manuscript; J.L.G. conception and design of research.