Calcineurin A-α suppression drives nuclear factor-κB-mediated NADPH oxidase-2 upregulation

Abstract

Calcineurin inhibitors (CNIs) are vital immunosuppressive therapies in the management of inflammatory conditions. A long-term consequence is nephrotoxicity. In the kidneys, the primary, catalytic calcineurin (CnA) isoforms are CnAα and CnAβ. Although the renal phenotype of CnAα^−/− mice substantially mirrors CNI-induced nephrotoxicity, the mechanisms downstream of CnAα are poorly understood. Since NADPH oxidase-2 (Nox2)-derived oxidative damage has been implicated in CNI-induced nephrotoxicity, we hypothesized that CnAα inhibition drives Nox2 upregulation and promotes oxidative stress. To test the hypothesis, Nox2 regulation was investigated in kidneys from CnAα^−/−, CnAβ^−/−, and wild-type (WT) littermate mice. To identify the downstream mediator of CnAα, nuclear factor of activated T cells (NFAT) and NF-κB regulation was examined. To test if Nox2 is transcriptionally regulated via a NF-κB pathway, CnAα^−/− and WT renal fibroblasts were treated with the NF-κB inhibitor caffeic acid phenethyl ester. Our findings showed that cyclosporine A treatment induced Nox2 upregulation and oxidative stress. Furthermore, Nox2 upregulation and elevated ROS generation occurred only in CnAα^−/− mice. In these mice, NF-κB but not NFAT activity was increased. In CnAα^−/− renal fibroblasts, NF-κB inhibition prevented Nox2 upregulation and reactive oxygen species (ROS) generation. In conclusion, these findings indicate that 1) CnAα loss stimulates Nox2 upregulation, 2) NF-κB is a novel CnAα-regulated transcription factor, and 3) NF-κB mediates CnAα-induced Nox2 and ROS regulation. Our results demonstrate that CnAα plays a key role in Nox2 and ROS generation. Furthermore, these novel findings provide evidence of divergent CnA isoform signaling pathways. Finally, this study advocates for CnAα-sparing CNIs, ultimately circumventing the CNI nephrotoxicity.

NEW & NOTEWORTHY A long-term consequence of calcineurin inhibitors (CNIs) is oxidative damage and nephrotoxicity. This study indicates that NF-κB is a novel calcineurin-regulated transcription factor that is activated with calcineurin inhibition, thereby driving oxidative damage in CNI nephropathy. These findings provide additional evidence of divergent calcineurin signaling pathways and suggest that selective CNIs could improve the long-term outcomes of patients by mitigating renal side effects.

INTRODUCTION

Calcineurin inhibitors (CNIs), such as cyclosporine A (CsA), tacrolimus, and voclosporin, are vital immunosuppressive therapies in the management of inflammatory conditions such as posttransplantation immunosuppression (1), lupus nephritis (2, 3), and rare cases of atopic dermatitis (4). Although CNIs have dramatically improved the quality of patient care, chronic immunosuppressant therapy has detrimental consequences. Specifically, patients experience various declines in renal function (5), and up to 15% eventually progress to end-stage kidney disease. Now, a new challenge becomes mitigating the side effects that stem from chronic immunosuppressant therapy without jeopardizing organs. As such, understanding the mechanisms underlying the nephropathy is paramount to the development of next-generation CNI immunosuppressants.

CNIs target calcineurin proteins to exert their immunosuppressive effects (6–8). Calcineurin proteins comprise a family of Ca⁺- and calmodulin-dependent phosphatases. These ubiquitously expressed phosphatases are notable for their key role in T cell function. Upon receiving an activation signal, the catalytic (CnA) subunit of calcineurin associates with the regulatory (CnB) subunit and calmodulin to form an active enzyme complex (1, 9). Nuclear factor of activated T cells (NFAT) proteins are well-recognized downstream calcineurin targets (10, 11). After dephosphorylation, NFAT proteins translocate into the nucleus and promote transcription of target genes including proinflammatory cytokines (12–14). By impeding calcineurin activity and subsequent downstream signaling, CNIs exert their immunosuppressive effects by preventing immune responses (1).

In the kidney, CnAα and CnAβ are the primary catalytic isoforms. CnAβ is the inducible isoform involved in renal pathophysiological responses (15, 16), whereas CnAα is the constitutively active isoform involved in kidney physiological processes (16–18). Regarding CNI-induced nephropathy, CnAα^−/− mice mirror features of CsA-treated mice unlike CnAβ^−/− mice (19). Although there is evidence that CnA isoforms are differentially regulated and possess distinct renal functions (15, 16, 18, 20), the isoform-specific signaling pathways involved in CNI nephropathy continue to be defined.

CNIs promote nephropathy, in part, by stimulating oxidative stress. In the kidney, NADPH oxidases (Nox) are the primary sources of reactive oxygen species (ROS). Nox enzymes are composed of several subunits that include catalytic, membrane-bound isoforms including Nox1–Nox5, as well as regulatory, cytosolic components including p67^phox, p47^phox, and p22 (21). Upon stimulation, subunits assembly into a functional complex that generates ROS by transferring an electron from NADPH to molecular oxygen. In pathological settings, Nox enzymes contribute to oxidative stress and renal damage (16, 22). Djamali et al. have provided strong evidence that Nox2 mediates CsA-induced nephropathy (23, 24). Specifically, this group demonstrated that CsA induces Nox2 upregulation and fibrosis, hallmarks of nephropathy. Notably, CsA-induced nephropathy is prevented in Nox2^−/− mice. Although Nox2 is identified as a critical mediator of CsA-induced pathological renal consequences, the specific CnA isoform involved and the sequence of events that precedes Nox2 regulation are unknown.

In this study, we aimed to identify the specific CnA isoform and downstream effectors involved in Nox2 regulation. As such, we tested the overarching hypothesis that CnAα negatively regulates Nox2-derived oxidative stress and that loss of CnAα function causes Nox2 upregulation. Using both in vivo and in vitro models of CNI-induced nephropathy, we demonstrated that 1) CnAα loss stimulates Nox2 upregulation, 2) NF-κB is a novel CnAα-regulated transcription factor, and 3) NF-κB mediates CnAα-induced Nox2 and ROS regulation. These novel results indicate that CnAα inhibition drives NF-κB activation, thereby promoting Nox2-mediated oxidative stress. This increased understanding of the initiating mechanisms underlying CNI nephropathy are vital to the development of next-generation CNI immunosuppressants that mitigate renal side effects.

EXPERIMENTAL DESIGN

CNI Models

In vivo.

Wild-type (WT) C57Bl/6 mice were purchased from Jackson Laboratory. Mice received intraperitoneal injections of CsA (20 mg/kg/day) or castor oil (Kolliphor EL; 1 mL/kg/day) for 4 wk, as previously described (25). All animal protocols and procedures were approved by the Animal Care and Use Committee of Emory University and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice were socially housed in standard cages with a regular 12:12-h light-dark cycle and fed standard rodent chow. Kidney analysis was performed on 12-wk-old male mice. Two separate studies were carried out with a total of four to seven mice examined. Representative images from one study are shown. Combined densitometric values from both studies are presented.

Global, constitutive CnAα^−/− and CnAβ^−/− mice were crossed with mice expressing an NFAT-responsive luciferase promoter, as previously described (15, 26). Experiments were carried out using these transgenic mice and littermate controls. All procedures were approved by the Atlanta Veterans Affairs Medical Center Institutional Animal Care and Use Committee and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice were socially housed in standard cages with a regular 12:12-h light-dark cycle and fed standard rodent chow. To prevent failure to thrive and early lethality, CnAα^−/− mice and littermates were supplemented with digestive enzyme-enriched chow to compensate for a salivary gland defect (15). Kidney analysis was performed on 6- to 8-wk-old male mice. Two separate studies were carried out with a total of five to six mice examined. Representative images from one study are shown. Combined densitometric values from both studies are presented.

In vitro.

Renal fibroblast cell lines were generated from the renal cortices of WT and CnAα^−/− male mice, as previously described (19). Cell lines were authenticated by IDEXX BioResearch to confirm species and sex. Cultured cells were grown at 37°C in 5% CO₂ in DMEM and F-10 Hams growth medium supplemented with 10% FBS, penicillin-streptomycin, and Plasmocin prophylactic, to prevent mycoplasma contamination. At 85% confluence, culture medium was changed to serum-free medium for 18–24 h. Afterward, cells were exposed to serum-free medium containing 10 nM caffeic acid phenethyl ester (CAPE) or 0.01% vehicle (ethanol) for 24 h.

Calcineurin Assessment

Calcineurin activity.

Calcineurin phosphatase activity was measured as previously described (16, 27). Briefly, cell lysates were incubated with the 5-carboxytetramethylrhodamine (TAMRA) fluorescently tagged calcineurin substrate-RII peptide. To remove phosphorylated RII peptide, the collected supernatant was incubated in a titanium oxide-coated plate (Glygen, Baltimore, MD). Afterward, the supernatant containing dephosphorylated RII peptide was collected and measured (540-nm excitation and 575-nm emission). The concentrations in experimental samples were extrapolated from a standard curve of purified calcineurin (Sigma-Aldrich).

ROS Measurement

Nox2 protein expression.

Western blot analysis was performed as previously described (16, 22). Membranes were immunoblotted with primary antibodies (1:1,000) specific for Nox2 (Abcam) and GAPDH. Densitometric analysis of immunoreactive bands were performed. Immunohistochemistry was performed to visualize Nox2 expression within kidneys. Paraffin-embedded kidney sections (5 µm thick) were processed and incubated with anti-Nox2 antibody (1:500). Images (×20 and ×60) of Nox2 fluorescence were acquired using a Keyence BZ-X800 Fluorescence Microscope.

Nox2 activation.

Nox2 and p67^phox association was examined by a coimmunoprecipitation assay, as we previously described (22). Briefly, protein lysates were incubated with protein A beads conjugated with p67^phox antibody (Cat. No. 07-002, Millipore). p67^phox-bound Nox2 protein was detected by Western blot analysis, as previously described (22).

ROS generation.

An amplex red (Invitrogen) assay was performed as previously described (16, 22, 28). Samples were incubated in amplex red and horseradish peroxidase. Fluorescence was measured (540-nm excitation and 590-nm emission) and compared with a standard curve to calculate H₂O₂ concentrations. H₂O₂ concentrations were normalized to cell number determined by measuring Hoechst fluorescence or tissue weight.

NFAT Activity Assessment

Renal NFAT activity.

Dissected renal cortices, outer medullae (OMs), and inner medullae (IMs) were homogenized, and NFATc-mediated luciferase activity was measured using a commercial kit (Promega, Madison, WI), as previously described (15). Briefly, lysates were incubated with luciferase assay reagents. Luminescence was measured using an OptoComp luminometer (MGM Instruments, Hamden, CT). Results were normalized by subtracting values obtained from identically processed samples from NFATc-luc negative littermate mice.

Renal NFAT DNA binding.

Active NFAT was qualified using a transcription factor array. Briefly, immobilized transcription factors were probed with whole kidney lysates. Immunoreactive bands were visualized and quantified. Densitometric values were normalized to control.

NF-κB Activity Assessment

NF-κB DNA-binding activity.

NF-κB (p65 and p50)-binding activity was assessed by a TransAM NF-κB Activation Assay kit (Active Motif), as described by the manufacturer. This ELISA-based assay measures binding of active NF-κB to its consensus site. Briefly, cells lysates were incubated with an oligonucleotide containing the NF-κB consensus site. Bound NF-κB was detected using primary antibodies that recognize the p65 or p50 subunits. NF-κB activity was determined by extrapolating the absorbance values of experimental samples from a standard curve of recombinant NF-κB proteins.

NF-κB activation.

Western blot analysis was performed as previously described (28). Membranes were immunoblotted with primary antibody (1:1,000) specific for NF-κB p100/p52 (Cell Signaling Technology).

Fibrosis Assessment

Fibronectin protein expression.

Western blot analysis was performed as previously described (19). Membranes were immunoblotted with primary antibody (1:1,000) specific for fibronectin (Abcam).

Collagen deposition.

Immunohistochemistry was performed to visualize fibrosis within kidneys. Paraffin-embedded kidney sections (5 µm thick) were processed and stained with trichrome blue. Images (×60) of extracellular collagen were acquired using a Keyence BZ-X800 Fluorescence Microscope.

Statistical Analysis

For all experiments, graphing and statistical analyses were performed using GraphPad software (Prism, San Diego, CA). Statistical tests were performed using either one-way ANOVA with a t test post test or two-way ANOVA with a Bonferroni’s post test to detect differences between experimental groups. Values of P < 0.05 were considered statistically significant.

RESULTS

CNI-Induced Renal Fibrosis Is Accompanied by Oxidative Stress and Nox2 Upregulation

We used an established mouse model of CNI-induced nephrotoxicity to investigate the mechanisms of renal oxidative stress. Compared with vehicle-treated mice, kidneys from CsA-treated mice had increased extracellular collagen deposition (Fig. 1A). Consistently, expression of the fibrotic protein fibronectin was significantly increased in CsA-treated mice [1.271 ± 0.0244 arbitrary units (AU)] compared with control mice (1.004 ± 0.0311 AU; Fig. 1B). Consistent with previously published findings (23, 24), CsA-induced ROS generation [vehicle: 1.000 ± 0.0628 relative light units (rlu)/µg protein vs. CsA: 1.828 ± 0.2059 rlu/µg protein; Fig. 1C] was associated with enhanced Nox2 activation as demonstrated by Nox2 and p67^phox association (vehicle: 1.000 ± 0.108 rlu/µg protein vs. CsA: 2.199 ± 0.117 rlu/µg protein; Fig. 1D). These CsA-induced renal effects were also associated with renal Nox2 protein upregulation (vehicle: 1.000 ± 0.200 rlu/µg protein vs. CsA: 3.07 ± 0.163 rlu/µg protein; Fig. 1D). Assessment of renal localization showed Nox2 expression within tubules but absence in glomeruli (Fig. 1E). Furthermore, compared with vehicle-treated mice, CsA treatment increased tubular Nox2 protein expression. Together, this model establishes that CNIs promote fibrosis and oxidative stress and enhance Nox2 expression and activation.

Figure 1.

Cyclosporine A (CsA)-induced renal fibrosis is accompanied by oxidative stress and NADPH oxidase-2 (Nox2) upregulation. To investigate renal consequences of calcineurin inhibitor treatment, kidneys from male C57Bl/6 mice treated with CsA (20 mg/kg/day) or vehicle for 4 wk were examined. A: extracellular collagen deposition was visualized by immunohistochemistry using trichrome blue (magnification: ×60). B: fibronectin protein abundance was quantified by Western blot. C: reactive oxygen species (ROS) levels were measured by amplex red. D: Nox2 activation was assessed by measuring Nox2 and p67^phox association via coimmunoprecipitation, and Nox2 protein abundance was quantified by Western blot. E: Nox2 protein was visualized by immunofluorescence (magnification: ×20 and ×60). Representative images are shown. n  =  4–7 mice/group. *P < 0.05 vs. vehicle-treated mice via one-way ANOVA. Circles, vehicle-treated mice; squares, CsA-treated mice.

Loss of CnAα Activity Promotes ROS Accumulation

The contribution of CnAα and CnAβ to basal calcineurin phosphatase activity and ROS regulation was investigated. To this end, kidneys from CnAα^−/−, CnAβ^−/−, and WT littermate mice were examined. The findings show that basal calcineurin activity was significantly reduced in CnAα^−/− kidneys (0.4075 ± 0.0830 ng/µg protein) compared with WT kidneys (0.8475 ± 0.0286 ng/µg protein; Fig. 2A). However, basal calcineurin activity was unaltered in kidneys from CnAβ^−/− mice (0.9275 ± 0.0735 ng/µg protein). In contrast, ROS levels were significantly increased in CnAα^−/− kidneys (1.766 ± 0.08864 rlu/µg protein) compared with WT kidneys (1.021 ± 0.0223 rlu/µg protein), whereas ROS levels were significantly reduced in CnAβ^−/− kidneys (0.5695 ± 0.0391 rlu/µg protein; Fig. 2B). Collectively, these findings provide evidence that CnAα is the renal calcineurin isoform that contributes to basal calcineurin phosphatase activity and is important for modulating basal ROS levels in kidneys.

Figure 2.

Catalytic calcineurin subunit A (CnA)α contributes to basal calcineurin activity and reactive oxygen species (ROS) generation. To investigate renal consequences of CnA isoform loss, kidneys from CnAα^−/− and CnAβ^−/− mice and wild-type (WT) littermates were examined. A: basal calcineurin activity was examined by a calcineurin phosphate assay. B: ROS levels were measured by amplex red. Amplex red samples were run in duplicate. n = 3–4 mice/group. *P < 0.05 vs. WT mice via one-way ANOVA. Circles, WT mice; squares, CnAα^−/− mice; triangles, CnAβ^−/− mice.

CnAα Regulates Nox2 Basally

Next, CnAα^−/− and CnAβ^−/− mice and WT littermates were used to identify the CnA isoform involved in Nox2 regulation. Consistent with increased ROS generation (Fig. 2B), Nox2 and p67^phox association, a maker of Nox2 activation, was significantly enhanced in CnAα^−/− kidneys (1.311 ± 0.0592 AU) compared with WT kidneys (1.046 ± 0.024 AU; Fig. 3A). However, Nox2 and p67^phox association was unaltered in CnAβ^−/− mice (0.8886 ± 0.1070 AU). Furthermore, compared with WT littermates (1.052 ± 0.0731 AU), renal Nox2 protein abundance (Fig. 3B) was significantly elevated in CnAα^−/− mice (2.490 ± 0.4381 AU) but not CnAβ^−/− mice (0.8291 ± 0.1518 AU). Similarly, renal Nox2 mRNA expression (Fig. 3C) was elevated in CnAα^−/− mice (1.527 ± 0.089 AU) but unaltered in CnAβ^−/− mice (0.862 ± 0.121 AU) compared with WT littermates (1.043 ± 0.139 AU). Although immunohistochemistry indicated lack of glomerular Nox2, increased tubular Nox2 expression was detected in CnAα^−/− mice but not CnAβ^−/− mice (Fig. 3D). These findings demonstrate the inverse relationship between CnAα activity and ROS levels. Specifically, these results show that CnAα suppresses Nox2 and that loss of CnAα activity stimulates Nox2 upregulation. Together, these findings provide support that CnAα is the calcineurin isoform that regulates Nox2 basally and is important for modulating basal ROS levels in kidneys.

Figure 3.

Catalytic calcineurin subunit A (CnA)α regulates NADPH oxidase-2 (Nox2) basally. To investigate renal consequences of CnA isoform loss on Nox2 regulation, kidneys from CnAα^−/− and CnAβ^−/− mice and wild-type (WT) littermates were examined. A: Nox2 activation was assessed by measuring Nox2 and p67^phox association via coimmunoprecipitation assays. B: Nox2 protein abundance was quantified by Western blot. C: Nox2 mRNA levels were quantified by quantitative RT-PCR. D: Nox2 protein was visualized by immunofluorescence (magnification: ×60). Representative images are shown. n = 5–6 mice/group. *P < 0.05 vs. WT mice via one-way ANOVA. Circles, WT mice; squares, CnAα^−/− mice; triangles, CnAβ^−/− mice.

CnAβ Selectively Regulates NFAT Activity

As a first approach to identify the downstream mediators of CnAα effects on Nox2, NFAT regulation by CnAα and CnAβ was investigated. Kidneys from CnAα^−/− and CnAβ^−/− mice and WT littermates expressing a NFAT luciferase responsive transgene were examined. Compared with WT mice, NFAT activity was unaltered in CnAα^−/− cortical (13.030 ± 2.186 vs. 13.096 ± 1.401 rlu/mg tissue), OM (37.458 ± 15.559 vs. 39.150 ± 5.462 rlu/mg tissue), and IM (8136.725 ± 237.105 vs. 8711.643 ± 686.706 rlu/mg tissue) nephron segments (Fig. 4A). Conversely, in CnAβ^−/− mice (Fig. 4B), NFAT activity was significantly reduced in cortical (7.014 ± 4.095 vs. 1.539 ± 0.250 rlu/mg tissue), OM (24.029 ± 2.006 vs. 2.975 ± 0.861 rlu/mg tissue), and IM (3,602.678 ± 186.796 vs. 33.604 ± 4.979 rlu/mg tissue) nephron segments compared with WT mice.

Figure 4.

Catalytic calcineurin subunit A (CnA)β selectively regulates nuclear factor of activated T cells (NFAT) activity. To investigate NFAT regulation by CnA isoforms, kidneys from CnAα^−/− and CnAβ^−/− mice and wild-type (WT) littermates were examined. A and B: NFAT activity in nephron segments was examined by measuring luciferase activity. C and D: NFAT and GATA DNA binding was quantified by a transcription factor array. n = 5–6 mice/group. *P < 0.05 vs. WT mice via one-way ANOVA. IM, inner medulla; OM, outer medulla. Squares, WT mice; triangles, CnAα^−/−and CnAβ^−/− mice.

To confirm isoform-specific NFAT regulation, NFAT DNA binding was assessed using a transcription factor array. Consistent with activity findings (Fig. 4, A and B), NFAT binding was not significantly different between CnAα^−/− (29.700 ± 2.829 AU) and WT mice (32.150 ± 0.779 AU; Fig. 4, C and D). However, in CnAβ^−/− mice, NFAT binding was reduced (7.950 ± 1.010 AU). Similar results were obtained for another well-characterized calcineurin substrate, GATA transcription factors (CnAα^−/−: 15.300 ± 0.866 AU, CnAβ^−/−: 2.650 ± 0.029 AU, and WT: 16.050 ± 0.318 AU). These findings demonstrate that loss of CnAβ but not CnAα suppresses NFAT activity. Furthermore, these results indicate that CnAβ is the primary calcineurin isoform responsible for NFAT regulation and that NFAT is not involved in Nox2 upregulation with CnAα inhibition.

CnAα Selectively Regulates NF-κB Activity

Since NF-κB is involved in Nox regulation (28), kidneys from CnAα^−/−, CnAβ^−/−, and WT mice were used to determine if NF-κB is a downstream target of CnAα. p65 and p100/p52 activities were examined by NF-κB activity assays. The findings show that renal p65 NF-κB activity (Fig. 5A) was reduced in CnAα^−/− mice (0.7366 ± 0.056 rlu/mg protein) compared with WT mice (1.000 ± 0.084 rlu/mg protein). In CnAβ^−/− mice, p65 NF-κB activity was unaltered (1.095 ± 0.075 rlu/mg protein). In contrast, p100/p52 NF-κB activity (Fig. 5B) was significantly enhanced in CnAα^−/− mice (2.661 ± 0.4682 rlu/mg protein) but reduced in CnAβ^−/− mice (0.419 ± 0.1038 rlu/mg protein) compared with WT littermates (1.062 ± 0.105 rlu/mg protein). To confirm p100/p52 activity, p52 abundance was assessed as a marker of p100 activation (Fig. 5C). Consistent with activity findings, renal p52 NF-κB abundance was increased in CnAα^−/− mice (2.190 ± 0.198 AU) but was not statistically different in CnAβ^−/− mice (1.368 ± 0.1097 AU) compared with WT mice (1.068 ± 0.1344 AU). Together, these findings indicate that CnAα is the primary calcineurin isoform responsible for NF-κB regulation. Furthermore, these results demonstrate that p100/p52 NF-κB is a downstream target of CnAα and that loss of CnAα results in NF-κB activation.

Figure 5.

Catalytic calcineurin subunit A (CnA)α selectively regulates NF-κB activity. To investigate NF-κB regulation by CnA isoforms, kidneys from CnAα^−/− and CnAβ^−/− mice and wild-type (WT) littermates were examined. A and B: p65 (A) and p52 (B) activity was examined by NF-κB activity assays. C: p52 protein abundance was quantified by Western blot. Representative images are shown. n = 4–6 mice/group. *P < 0.05 vs. WT mice via one-way ANOVA. Circles, WT mice; squares, CnAα^−/− mice; triangles, CnAβ^−/− mice.

NF-κB Mediates CnAα-Induced Nox2 Regulation

Finally, to determine if NF-κB is involved in CnAα-mediated Nox2 regulation, CnAα^−/− and WT renal fibroblasts were treated with CAPE (a NF-κB inhibitor) or vehicle. Consistent with in vivo findings (Fig. 3C), Nox2 mRNA levels were elevated in CnAα^−/− renal fibroblasts (3.320 ± 0.335 AU) compared with WT cells (1.000 ± 0.035 AU; Fig. 6A). In WT cells, NF-κB inhibition reduced basal Nox2 mRNA expression (0.390 ± 0.017 AU) compared with vehicle-treated renal fibroblasts (1.000 ± 0.035 AU). CAPE treatment also attenuated Nox2 mRNA in CnAα^−/− cells (1.647 ± 0.050 vs. 3.320 ± 0.335 AU). Similarly, NF-κB inhibition by CAPE suppressed Nox2 protein expression (Fig. 6B) in both WT (0.521 ± 0.107 vs. 1.056 ± 0.006 AU) and CnAα^−/− renal fibroblasts (0.662 ± 0.113 vs. 1.870 ± 0.308 AU). Consistent with reduced Nox2 expression, CAPE treatment significantly reduced ROS levels in CnAα^−/− cells (0.862 ± 0.096 vs. 1.513 ± 0.117 µM/DNA; Fig. 6C). Collectively, these results demonstrate that NF-κB mediates CnAα-induced Nox2 regulation. Altogether, this study indicates that loss of CnAα and subsequent NF-κB activation promotes Nox2 upregulation.

Figure 6.

NF-κB mediates catalytic calcineurin subunit A (CnA)α-induced NADPH oxidase-2 (Nox2) regulation. To determine if NF-κB plays a role in CnAα-mediated Nox2 regulation, renal-derived CnAα^−/− and wild-type (WT) fibroblasts were treated with caffeic acid phenethyl ester (CAPE), a NF-κB inhibitor. A: Nox2 mRNA levels were quantified by quantitative RT-PCR. B: Nox2 protein abundance was quantified by Western blot. C: reactive oxygen species (ROS) levels were measured by amplex red. Representative Western blots are shown. Amplex red samples were run in duplicate. n = 3 independent experiments. *P < 0.05 vs. vehicle-treated WT fibroblasts; #P < 0.05 vs. vehicle-treated CnAα^−/− fibroblasts via two-way ANOVA. Circles, vehicle-treated fibroblasts; squares, CAPE-treated fibroblasts.

DISCUSSION

CNIs promote nephropathy, in part, by stimulating renal oxidative damage (29, 30). Since CnA isoforms are differentially regulated and possess distinct functions, identifying the specific CnA isoform involved is critical in informing the development of next-generation CNIs. Gooch et al. demonstrated that loss of CnAα recapitulates CsA nephropathy, whereas loss of CnAβ does not alter the renal phenotype (19). Additionally, Djamali et al. demonstrated that Nox2 mediates CsA nephropathy (23, 24). To determine if Nox2 and CnAα cooperate in a common pathophysiological pathway, we tested the overarching hypothesis that CnAα negatively regulates Nox2 and that loss of CnAα function causes Nox2 upregulation. Using both in vivo and in vitro models of CNI-induced nephropathy, we demonstrated that 1) CnAα loss stimulates Nox2 upregulation, 2) NF-κB is a novel CnAα-regulated transcription factor, and 3) NF-κB mediates CnAα-induced Nox2 and ROS regulation. These results indicate that CnAα inhibition drives NF-κB activation, thereby promoting Nox2-mediated oxidative stress.

CnAα Loss Stimulates Nox2 Upregulation

Consistent with CnAα negatively regulating basal ROS levels, its absence in CnAα^−/− mice results in increased ROS accumulation, whereas its presence in CnAβ^−/− mice is sufficient to reduce ROS levels (Fig. 2B). These findings demonstrate that CnAα is the basally active renal isoform and that loss of its activity promotes oxidative stress. Djamali et al. demonstrated that the ROS-generating enzyme Nox2 mediates CsA nephropathy (23, 24). Specifically, this group showed that loss of Nox2 prevents CsA-induced renal oxidative damage and fibrosis, indicating that the Nox2 isoform plays a critical role in CsA nephropathy. Additionally, our study used CsA as a model of CNI nephrotoxicity. However, due to its better efficacy and milder range of adverse effects, tacrolimus largely outpaces CsA in clinical success. As such, future studies are needed to specifically investigate the underlying mechanisms responsible and whether Nox2/ROS is a common pathway for both CsA- and tacrolimus-induced nephrotoxicity. The similarities and differences between these widely used immunosuppressive agents have been previously discussed (1, 8, 31, 32).

Although this and other study’s findings are consistent with calcineurin functioning as a negative regulator of Nox2, the specific CnA isoform involved remained unknown. Our data demonstrate that loss of CnAα but not CnAβ stimulates Nox2 expression and activation (Fig. 3). These findings indicate that the CnAα isoform functions as a negative regulator of Nox2. Therefore, specific loss of CnAα activity initiates pathophysiological events that results in Nox2 upregulation, ROS generation, and subsequent oxidative damage (Fig. 7).

Figure 7.

NF-κB is a novel catalytic calcineurin subunit A (CnA)α-regulated transcription factor that is activated with CnAα inhibition, thereby driving NADPH oxidase-2 (Nox2)-mediated oxidative damage in calcineurin inhibitor (CNI) nephropathy. CnAα loss or inhibition (1) activates the transcription factor NF-κB and unleashes a wave of NF-κB targets, including Nox2 (2). The resultant enhanced Nox2 protein expression and subsequent activation (3) leads to increased reactive oxygen species (ROS) generation (4). These events culminate in oxidative damage (5) that contributes to nephropathy. [Image was created with BioRender.com.]

NF-κB Is a Novel CnAα-Regulated Transcription Factor

NFAT is a well-characterized target of calcineurin (10, 11, 26) and was a logical candidate to mediate regulation of Nox2 by CnAα. However, our data show that loss of CnAβ but not CnAα reduces NFAT transcriptional activity (Fig. 4), consistent with our previous in vitro work showing that NFAT is a CnAβ target (16). Consistently, NFAT mediates CnAβ modulation of the immune response. Studies distinguishing the roles of CnAα and CnAβ in T cell activation have been previously discussed (1, 31, 32).

It is well established that NF-κB regulates Nox enzymes (28, 33), leading us to consider that NF-κB may be downstream of CnAα. Notably, our data demonstrate that loss of CnAα selectively stimulates activation and activity of the noncanonical NF-κB pathway (Fig. 5), suggesting that CnAα is an endogenous inhibitor of NF-κB. Interestingly, cyclooxygenase-2 (Cox-2) is a NF-κB-regulated enzyme negatively impacted by CNIs (34). Although Cox-2 is also involved in Nox regulation, it is unclear whether CnAα and Cox-2 function in a common pathway to modulate Nox2 and is a needed focus for future studies.

The role of the canonical and noncanonical NF-κB pathways has been well studied in cells of the immune system. In immune cells, NF-κB uses both pathways to not only produce prooxidants that provide the cells the ability to kill foreign pathogens but also antioxidants that protect these cells against excessive oxidative stress (35). Specifically, Nox2 expression is induced by canonical NF-κB signaling in phagocytes to mount a rapid initial defense against invading microorganisms (36). In contrast, the noncanonical NF-κB pathway is activated to stimulate a prolonged response (37). In this study, CnAα inhibition-induced noncanonical NF-κB signaling is suggestive of a prolonged cell response. With fibrosis well characterized as a chronic state, it is conceivable that activation of the noncanonical NF-κB pathway maintains this long-term process. Although our data do not show activation of the canonical NF-κB pathway, it is quite possible that the chronic CnAα loss in CnAα^−/− mice prevents detection of this early, rapid response. Furthermore, the exact roles of both the canonical and noncanonical NF-κB signaling pathways in the modulation of various responses in the kidneys remain to be better understood.

NF-κB Mediates CnAα-Induced Nox2 and ROS Upregulation

Since resident fibroblasts are key players that drive nephropathy (38, 39), we used immortalized, resident renal fibroblasts isolated from CnAα^−/− mice and WT littermates to directly place CnAα in the NF-κB/Nox2 pathway. Specifically, our findings demonstrate that loss of CnAα promotes Nox2 upregulation (Fig. 6). However, pharmacological inhibition of NF-κB transcriptional activity prevents Nox2 upregulation. These novel findings uncover that NF-κB mediates CnAα-induced Nox2 regulation. Furthermore, NF-κB represents a novel mechanism by which CnAα regulates oxidative stress. Importantly, these findings fill a gap in knowledge by identifying the divergent renal signaling mechanisms that contribute to distinct CnA isoform functions.

The present study adds an important new mechanism to our understanding of CNI nephropathy: loss or inhibition of CnAα upregulates the transcription factor NF-κB and unleashes a wave of NF-κB targets including Nox2 (Fig. 7). In addition to the pathological response demonstrated in this study, other NF-κB targets, such as Cox2, inducible nitric oxide synthase, and tumor necrosis factor, may be involved in CNI nephrotoxicity and represent an important new area for investigation. Expanding understanding of CNI pathological mechanisms is an important outcome of this study. Critically, we have separated the intended effects of CNIs (immunosuppression via CnAβ inhibition) from the deleterious CNI renal effects (ROS accumulation via CnAα inhibition). Additionally, this study provides supporting evidence to the established finding that a loss of CnAα promotes renal damage, an effect absent with CnAβ inhibition. Since CnAβ plays a critical role in T cell-mediated immune responses (40, 41), a renal-sparing drug will likely be one that does not substantially reduce CnAα activity and thereby promote NF-κB activation and Nox2 upregulation. Thus, identification of these specific CnA isoform mechanisms have an important impact in the development of newer generation CNIs that prevent the associated renal complications while maintaining immunosuppression.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R21DK119879 (to C.R.W.) and American Heart Association Grant 16SDG27080009 (to C.R.W.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.M.C., J.L.G., C.R.W., and H.P.M. conceived and designed research; A.M.C., C.R.W., A.C.U., C.E.F., K.N.K., and V.A.L. performed experiments; A.M.C., J.L.G., C.R.W., C.E.F., and K.N.K. analyzed data; A.M.C., J.L.G., C.R.W., C.E.F., R.S.H., and H.P.M. interpreted results of experiments; A.M.C., C.R.W., and C.E.F. prepared figures; A.M.C., J.L.G., C.R.W., and C.E.F. drafted manuscript; A.M.C., J.L.G., C.R.W., A.C.U., C.E.F., K.N.K., V.A.L., Y.B., H.C., R.S.H., and H.P.M. edited and revised manuscript; A.M.C., J.L.G., C.R.W., A.C.U., C.E.F., K.N.K., V.A.L., Y.B., H.C., R.S.H., and H.P.M. approved final version of manuscript.