Cytochrome-Derived EETs, VEGF-A, and NOX4: Piecing the Puzzle Together

In this issue of Diabetes, Njeim et al. (1) describe, for the first time, vascular endothelial growth factor A (VEGF-A) as a link between reduced cytochrome P450 isoform–derived renoprotective epoxyeicosatrienoic acids (EETs) and increased NADPH oxidase 4 (NOX4)–derived formation of reactive oxygen species (ROS) in a rat model of type 1 diabetes–associated kidney disease.

The group has previously demonstrated that this cytochrome P450–dependent system plays a key role in proximal tubular injury and fibrosis in diabetes in rats (2–4). The authors showed that high glucose increased ROS production and altered expression of cytochrome P450 4A (CYP4A) and cytochrome P450 2C (CYP2C), which increased production of deleterious 20-hydroxyeicosatetraenoic acids (20-HETEs) and reduced bioavailability of protective EETs (2–4). Both 20-HETEs and EETs have been shown to play a role in hypertension, but their importance in diabetes and kidney disease is less clear. 20-HETEs and EETs are both products of arachidonic acid metabolism and can play opposing roles depending on disease type, disease state, and type of tissue. Whereas CYP4A is upregulated in diabetes and produces increased amounts of 20-HETEs, cytochrome P450 2C11 (CYP2C11) is downregulated, leading to reduced EET formation and bioavailability in diabetic kidney disease (DKD).

Furthermore, the authors have shown that the upregulation of HETEs is associated with increased NOX4 expression and activation (2–4); however, the relationship between EETs and NOX4 and VEGF-A is not clear.

As detailed in their article in this issue of Diabetes, Njeim et al. (1) used a soluble epoxide hydrolase inhibitor that hydrolyzes EETs into less active diols. By blocking the inactivating enzyme, the authors achieved an increase in the tissue concentration of renoprotective EETs. 12-(3-Adamantan-1-yl-ureido)-dodecanoic acid (AUDA) increased EET bioavailability and reduced ROS and NOX4 expression as well as NADPH oxidase activity. This was associated with renoprotective effects, including the reduction of albuminuria and kidney fibrosis and, specifically, podocyte protection by reducing podocyte depletion and apoptosis and restoring nephrin expression.

Downregulation of CYP2C11-derived EET formation in diabetes was correlated with activation of the VEGF-A signaling pathway. AUDA treatment of diabetic rats for 8 weeks also attenuated the diabetes-induced increase in VEGF-A expression.

As the next step, the authors aimed to block the VEGF-A pathway via two pharmacological approaches. First, they used SU 5416, a selective tyrosine kinase inhibitor that blocks the activating VEGF receptor autophosphorylation. Second, they used bevacizumab, a humanized anti-VEGF monoclonal antibody administered for 8 weeks starting after 5 weeks of diabetes in rats.

Both strategies to block VEGF-A attenuated kidney injury, including fibrosis and podocyte apoptosis, and restored nephrin expression, which was also associated with inhibition of NOX4 on the gene and protein levels as well as attenuation of NOX-derived ROS formation. The in vivo studies were complemented and confirmed by in vitro studies, including studies in human podocytes. In vitro, silencing of VEGF-A by siRNA replicated the findings obtained with the pharmacological approaches to inhibit VEGF-A in the rat model of DKD.

Although the authors did not show that VEGF-A inhibition directly restored EET levels, they provided conclusive in vivo evidence that the diabetes-induced downregulation of CYP2C-mediated EET formation is associated with upregulation of renal VEGF-A as well as of NOX4 and its derived ROS, thereby suggesting that VEGF-A could be a mechanistic link between EETs and NOX4.

Increased VEGF expression in the diabetic kidney was first described by Cooper et al. (5). Using radioligand binding assays and immunohistochemistry, the authors showed that VEGF is predominantly expressed in podocytes but also in distal tubules and collecting ducts. The VEGF receptor 2 (VEGFR-2) resides predominantly in endothelial cells and is upregulated early and in a transient manner in the course of diabetes (5). Podocytes and proximal tubular cells express VEGF-A throughout life.

VEGF is a pleiotropic protein excreted by podocytes, and it acts in a paracrine and autocrine manner as a survival factor for endothelial cells, podocytes, and mesangial cells. VEGF-A regulates the slit-diaphragm integrity, nephrin expression, and podocyte structure via VEGFR-2–nephrin–Nck–actin interactions. Increased VEGF-A expression, as observed in diabetes, is associated with low nitric oxide bioavailability (6). Podocytes produce VEGF-A by alternative splicing. Some isoforms are excreted, and others bind to the receptors VEGFR-1 and VEGFR-2 and neuropilin 1 and 2 coreceptors. Furthermore, hypoxia is also a strong inducer of VEGF-A gene expression, which is considered a hypoxia-inducible gene and promoter of angiogenesis (7).

VEGF-A expression is increased by high glucose and has been shown to play an important role in DKD. However, VEGF-A has been a difficult target to treat in various diseases. Physiological VEGF-A signaling is essential. Genetic and embryonic deletion of VEGF-A or its receptors is lethal. VEGF-A is increased in various diseases and is mostly deleterious, but it also regulates neoangiogenesis in cardiovascular disease by regulation of neoangiogenesis in ischemia (8).

The development of hypertension and thrombrotic microangiopathy resulting in glomerular disease has been reported following a systemic VEGF blockade and also following intravitreal use of VEGF inhibitors in the treatment of diabetic retinopathy (9). Studies in mice suggest that podocyte-derived VEGF is essential in maintaining a healthy glomerular endothelium. Therefore, it is not easy to achieve the optimal level of inhibition of VEGF–VEGFR signaling in disease, including in diabetes.

VEGF-A has been shown to induce NOX4 expression and activity. On the other hand, we and others have shown that NOX4-derived ROS formation activates VEGF-A expression and signaling (10,11).

There is now a plethora of evidence for the role of NOX4 in kidney diseases and other fibrotic diseases, including those affecting the lung and liver (12). NOX4 is rapidly upregulated by high glucose, transforming growth factor-β, and other factors in the diabetic milieu (13). Genetic deletion of NOX4 has been shown to attenuate albuminuria, renal fibrosis, and inflammation, to reduce glomerular VEGF expression, and to promote podocyte protection (10). A podocyte-specific deletion of NOX4 was associated not only with decreased VEGF expression in glomeruli but also with reduced urinary VEGF-A excretion and reduced renal fibrosis and mesangial expansion in a mouse model of type 1 DKD (11). In other studies, the global deletion of NOX4 reduced mitochondrial and cytosolic ROS production in association with attenuation of the protein kinase C pathway and VEGF-A reduction (14).

Inhibition of ROS generation is also a challenge. Most NOX inhibitors are not specific. However, more recently, specific NOX inhibitors have been developed. One of them is the NOX1/4 inhibitor GKT137831 (setanaxib), which has been shown to reduce kidney disease and to be vasculoprotective by inhibiting NOX1 and NOX4 (15). The renoprotective effects of pharmacological inhibitors of NOX4 have been shown in various models of type 1 and type 2 DKD (16,17). Indeed, the NOX1/4 inhibitor GKT 137831 has also been shown to reduce VEGF-A expression (10,15). Furthermore, pharmacological NOX inhibition has conferred podocyte protection by reducing podocyte apoptosis and restoring nephrin expression, but the effects on EET levels have not been investigated yet.

There is clear evidence for a close association between the NOX4 and the VEGF-A pathways. NOX4 promotes the directed migration of endothelial cells by stabilizing (VEGFR-2) protein in the endoplasmic reticulum, and it maintains stable VEGFR-2 expression levels at the surface (18). NOX4-derived H₂O₂ in part stimulates NOX2 to increase mitochondrial ROS via pSer36-p66Shc, thereby increasing VEGFR-2 signaling and angiogenesis in endothelial cells (19). NOX4-derived H₂O₂ also stimulates endothelial progenitor cells (20).

Based on the findings of Njeim et al. (1), in this article we propose not only that VEGF-A stimulates NOX4 expression and activation in a unidirectional manner but also that NOX4 upregulates VEGF-A, suggesting a bidirectional forward activation loop (Fig. 1).

Whether NOX4 inhibition or VEGF-A inhibitors restore EET levels and if and how this contributes to renoprotective and podocyte protective effects is not clear yet.

The question arises as to how to best target the EET–VEGF-A–NOX4 axis in diabetes. As outlined above, directly targeting VEGF-A is difficult. The clinical use of soluble epoxide hydrolase has not been proven and is limited by side effects on angiogenesis (21). Given that VEGF-A and NOX4 induce each other in a bidirectional manner, pharmacological targeting of NOX4 may be the better and more feasible approach to target all contributing pathways given that NOX4 inhibitors are now in clinical development. Furthermore, 20-HETEs and EETs are known to signal through G protein–coupled receptors (GPCRs), and given that podocytes express dozens of GPCRs (22–25), including orphan GPCRs (26), this pathway might represent an alternative way to target the EET–VEGF-A–NOX4 axis more selectively.

In summary, the article by Njeim et al. provides new evidence for a link between cytochrome P450–derived EETs, VEGF-A, and NOX4 in DKD. There are some limitations to this study. Other EETs and 20-HETEs than the ones investigated here may also play a role. Inhibitors may not be specific. It is difficult to directly translate these findings to the human context. The treatment is initiated after 5 weeks of diabetes, which represents an earlier stage of DKD. Whether this approach is also effective in a more advanced stage of disease remains to be shown.

Nevertheless, the findings provide a new avenue to pursue novel targets linked to the cytochrome P450–EET–VEGF-A–NOX4 pathway in DKD. Further studies are needed to investigate this in detail and to identify the best inhibitory approach as well as the ideal dosing regimen.