Mineralocorticoid Receptor in Endothelial Cells Contributes to Vascular Endothelial Growth Factor Receptor Inhibitor-Induced Vascular and Kidney Damage

Abstract

BACKGROUND

Vascular endothelial growth factor receptor inhibitors (VEGFRis) improve cancer patient survival by inhibiting tumor angiogenesis. However, VEGFRis induce treatment-limiting hypertension which has been associated with impaired vascular endothelial cell (EC) function and kidney damage. The mineralocorticoid receptor (MR) regulates blood pressure (BP) via its effects on the vasculature and the kidney. Thus, we interrogated the role of the MR in EC dysfunction, renal impairment, and hypertension in a mouse model of VEGFRi-induced hypertension using sorafenib.

METHODS

EC dysfunction in mesenteric arterioles was assessed by immunoblotting for phosphorylation of endothelial nitric oxide synthase (eNOS) at serine 1177. Renal damage was measured by assessing glomerular endotheliosis histologically. BP was measured using implanted radiotelemetry.

RESULTS

Six days of sorafenib treatment significantly impaired mesenteric resistance vessel EC function, induced renal damage, and increased BP. Pharmacologic MR blockade with spironolactone prevented the sorafenib-induced decline in eNOS phosphorylation and renal glomerular endotheliosis, without affecting systolic BP (SBP) or diastolic BP. Mice with the MR knocked out specifically in ECs (EC-MR-KO) were protected from sorafenib-induced EC dysfunction and glomerular endotheliosis, whereas smooth muscle cell-specific MR (SMC-MR) knockout mice were not. Neither EC-MR nor SMC-MR knockout affected the degree to which sorafenib increased SBP or diastolic BP.

CONCLUSIONS

These results reveal that the MR, specifically in EC but not in SMCs, is necessary for VEGFRi-induced renal and vascular injury. While ineffective at lowering SBP, these data suggest potential therapeutic benefits of MR antagonists, like spironolactone, to protect the vasculature and the kidneys from VEGFRi-induced injury.

Graphical Abstract

Graphical Abstract.

Vascular endothelial growth factor (VEGF) signaling is upregulated in tumors and promotes angiogenesis, thereby fostering tumor growth and metastasis.¹ As such, VEGFR inhibitors (VEGFRis) significantly improve cancer patient survival.² However, VEGFRis induce cardiovascular toxicity that limits cancer treatment and survival.³ The most common side effect of VEGFRis is increased blood pressure (BP), which results in new or exacerbated hypertension in up to 80% of patients.² The precise molecular mechanism of VEGFRi-induced hypertension remains unclear but a prevailing notion in the field is that vascular endothelial cell (EC) dysfunction leads to increased vasomotor tone that raises BP.³ Additionally, VEGFRis induce nephrotoxicity, characterized by proteinuria and glomerular endotheliosis, which may also contribute to or exacerbate existing hypertension.⁴

The mineralocorticoid receptor (MR) is a steroid hormone receptor that regulates renal electrolyte balance and hence BP in response to the hormone aldosterone.⁵ MR activation in the kidney promotes sodium and water reabsorption in the distal nephron and glomerular filtration and barrier function in the proximal nephron. In addition to the kidney, the MR is expressed in vascular ECs and smooth muscle cells (SMCs).⁵ Vascular MRs contribute to BP by regulating blood vessel diameter and tone and promoting vessel inflammation and remodeling associated with hypertension and heart disease.⁵ There are three MR antagonists – spironolactone, eplerenone, and finerenone – which are approved to treat heart failure, chronic kidney disease, and hypertension, resulting in improvements in cardiovascular health even at doses that do not alter BP.⁵

The resistance vasculature is an important determinant of systemic vascular resistance and thus BP. ECs produce factors that promote vessel constriction or relaxation while SMCs mediate the changes in vessel diameter in response to these factors to control myogenic tone and vascular resistance. Nitric oxide (NO) is a potent EC-derived vasodilator that diffuses to SMCs resulting in SMC relaxation and vessel dilation.⁶ VEGF stimulation promotes the phosphorylation of endothelial nitric oxide synthase (eNOS) at serine 1177 (p-eNOS), which activates eNOS to produce more of the vasodilator NO.⁶ Conversely, VEGFRis induce vascular endothelial dysfunction characterized by decreased p-eNOS levels and lower NO bioavailability.⁶ MR in ECs and SMCs also contributes to resistance vessel function.⁷ Specifically, activation of MR in ECs contributes to decreased p-eNOS in several disease states and MR in SMCs regulates myogenic tone with aging.⁷ However, whether MR is involved in VEGFRi-induced EC dysfunction or hypertension has not been investigated.

The kidney also plays an essential role in BP regulation. The primary filtration unit of the kidney is the renal glomerulus, which is composed of fenestrated ECs that also depend on VEGF and NO signaling for normal function.⁸ VEGF receptor inhibition in patients with cancer results in glomerular endotheliosis characterized by swelling of glomerular ECs and occlusion of glomerular capillaries.⁸ Glomerular endotheliosis leads to proteinuria, impaired pressure natriuresis, renal dysfunction, and hypertension in preeclampsia as well as with VEGFRi treatment.⁸ We recently demonstrated that SMC-MR plays a role in the enhanced hypertensive response after exposure to a preeclampsia model induced by VEGF inhibition,⁹ but the role of SMC-MR in VEGFRi-induced hypertension has not been tested.¹⁰

Thus, given the importance of MR in regulating vascular health and BP via its impact on the vasculature and the kidney, we investigated whether MR plays a role in the pathophysiological sequelae of VEGFRi treatment. Using a mouse model of VEGFRi-induced hypertension caused by sorafenib treatment, we examined the impact of systemic pharmacologic MR blockade with spironolactone vs. placebo and compared this to genetic MR deletion specifically in SMCs or ECs. The impact on VEGFRi-mediated decline in resistance vessel p-eNOS, induction of renal glomerular endotheliosis, and hypertension were measured. Overall, MR blockade attenuated VEGFRi-induced vascular and renal dysfunction without impacting the rise in BP. The benefits of MR blockade were reproduced by EC-specific MR knockout, but not SMC-specific MR deletion, in mice.

METHODS

Sorafenib-induced hypertension model in mice

All mouse studies were approved by the Tufts University Institutional Animal Care and Use Committee and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Male mice were housed in standard cages with a 12-hour light–dark cycle and allowed continuous access to standard food and water.¹¹ C57BL/6J mice (Jackson Laboratory, 3–4 months old) were implanted with placebo or spironolactone pellets (21-day timed release; 20 mg/kg/day; Innovative Research of America) subcutaneously as described.¹¹ Inducible SMC-specific MR knockout (SMC-MR-KO, MR floxed/SMA-Cre-ERT2)¹¹ mice and EC-specific MR knockout (EC-MR-KO, MR floxed/constitutive VECad-Cre)¹² and Cre-negative littermates were generated by methods we previously described and extensively validated.^11,12 A cohort of each type of mice was implanted with radiotelemetry catheters (TA11PA-C10, Data Sciences International) via the carotid artery into the aortic arch under isoflurane anesthesia with buprenorphine as postoperative analgesia (0.1 mg/kg subcutaneous injection), as described.¹¹ All mice were allowed to recover from surgery for 7 days. Systolic blood pressure (SBP) was then measured for 30 seconds every hour. SBP was measured for 3 days prior to starting sorafenib treatment and this defined the baseline SBP (day 0) for each mouse. Mice were then administered vehicle (10% dimethyl sulfoxide (DMSO), 40% PEG-300, 10% Tween-20 in normal saline) or sorafenib (MedChem Express, 60 mg/kg/day) via daily oral gavage (0.2 ml) for 5 days. The 24-hour average SBP was reported for each subsequent day for each mouse. At the end of the study, mice were euthanized under isoflurane anesthesia and mesenteric vessels and kidneys were flash frozen in liquid nitrogen and stored at −80 °C.

Immunoblotting

Mesenteric vascular arcades were combined from 2 mice and whole vessel lysate probed with primary antibodies (phospho-eNOS serine 1177 (BD Bioscience, #612393, 1:1,000), total eNOS (BD Bioscience, #610297, 1:1,000), endothelin-1 (ET-1; Abcam, #ab117757, 1:500), and GAPDH (Cell Signaling, #97166, 1:2,000) overnight followed by species-specific anti-rabbit horseradish peroxidase secondary antibodies (Cell Signaling, 1:1,000) for 1 hour, and imaged with ECL reagent (Thermo Fisher) and quantified using Fiji software as described.¹³

Kidney histology

Mouse kidneys were fixed in 10% neutral buffered formalin (Sigma), embedded in paraffin and 5-µm sections were stained with hematoxylin and eosin. Glomerular endotheliosis was scored in a blinded fashion based on an established 3-point glomerular endotheliosis scoring system with 0 = normal, 0.5 = 1%–9%, 1 = 10%–19%, 1.5 = 20%–50%, 2 = 50%–79%, and 3 = 80%–100% occlusion of the glomerular vessels, as previously described.¹⁴ Fifteen to twenty glomeruli per kidney were scored and the values averaged for each mouse.

Statistics

All statistical analyses were performed using GraphPad Prism version 10.2.0. For BP data, a 2-way repeated measures ANOVA with Holm–Sidak post-test was employed. All other data were analyzed by 1-way ANOVA with Holm–Sidak post-test or 2-way ANOVA with Fisher’s uncorrected LSD post-test, as indicated in each figure legend. All error bars express mean ± SEM.

RESULTS

MR blockade protects against VEGFRi-induced renal dysfunction without preventing hypertension

The impact of pharmacologic MR inhibition with spironolactone on VEGFRi-induced resistance vessel eNOS dysfunction, renal glomerular endotheliosis, and BP elevation was first examined (Figure 1a–c). Resistance vessel eNOS activity (serine 1177 phosphorylated (p-eNOS) relative to total eNOS) decreased significantly in mice treated with sorafenib vs. vehicle, and this decrease was significant only in those mice with a placebo pellet (Figure 1a). In mice treated with spironolactone, sorafenib did not significantly impact eNOS activity. Of note, in the absence of sorafenib, spironolactone significantly decreased the level of p-eNOS/total eNOS compared with those with a placebo pellet, suggesting that MR inhibition decreased basal eNOS activity. Neither sorafenib nor spironolactone treatment significantly modified ET-1 protein levels in mesenteric resistance vessels (Figure 1a, right). Renal glomerular endotheliosis was also significantly induced in mice treated with sorafenib compared with vehicle, and this increase was prevented in mice co-treated with spironolactone (Figure 1b). As such, sorafenib-treated mice co-treated with spironolactone had significantly less glomerular endotheliosis compared with placebo co-treated mice. In mice with implanted radiotelemetric BP monitors, this spironolactone dose did not significantly alter SBP during the 3-day interval before initiating sorafenib treatment (Figure 1c). Sorafenib treatment induced a significant and sustained increase in SBP (subject: P < 0.0001; time: P < 0.0001; Figure 1c) with no difference between mice treated with placebo and spironolactone. Thus, in the context of VEGFRi treatment, global MR inhibition was protective against renal impairment and attenuated resistance vessels eNOS dysfunction with no impact on the degree of BP elevation. Also, MR inhibition decreased basal eNOS activity in healthy mice without VEGFRi treatment.

Figure 1.

Mineralocorticoid receptor (MR) antagonism and MR deletion from endothelial cells (ECs), but not smooth muscle cells (SMCs), protects from VEGF receptor inhibitor-induced endothelial dysfunction and renal glomerular endotheliosis, but not from hypertension. Mice were treated with vehicle (veh) or sorafenib (sor) for 5 days to induce an increase in systolic blood pressure (BP); (a–c) compares mice treated with placebo or the MR antagonist spironolactone (Spiro), (d–f) compares SMC-MR knockout mice to MR-intact littermate controls, and (g–i) compares EC-MR knockout mice to MR-intact littermate controls. (a, d, g) Endothelial dysfunction measured in mesenteric resistance vessels: representative immunoblots of mesenteric vessel tissue lysate and quantifications of endothelial nitric oxide synthase phosphorylation on the activation site (pS1177-eNOS) normalized to total eNOS and endothelin 1 (ET-1) normalized to GAPDH. N = 4–5 pooled samples (8–10 mice) per group. Two-way ANOVA with Fisher’s uncorrected LSD post-test. (b, e, h) Renal glomerular endotheliosis: representative histologic images of H&E-stained glomeruli and quantification of the glomerular endotheliosis score in 15–20 glomeruli per kidney. Black-outlined triangles indicate representative patent glomerular capillaries. Scale bar = 30 µm. N = 6–13 mice/group. Two-way ANOVA with Fisher’s uncorrected LSD post-test. (c, f, i) VEGFRi-induced increase in systolic BP (SBP) measured by implanted radiotelemetry. For (f) and (i), day 0 = 3-day average SBP before sorafenib treatment. The dotted vertical line denotes the start of sorafenib treatment. N = 6–9 mice/group. Two-way ANOVA with repeated measures and Holm–Sidak post-test. There was no significant difference in BP by treatment or genotype. Abbreviations: Gt, genotype; Intx, interaction; SMC-MR, smooth muscle cell specific; VEGFRi, vascular endothelial growth factor receptor inhibitors. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

SMC-MR contributes to basal eNOS activity but is not necessary for VEGFRi-induced resistance vessel dysfunction, renal damage, or hypertension

SMC-MR knockout mice were treated with sorafenib and resistance vessel eNOS activity, glomerular endotheliosis, and SBP were compared with MR-intact littermates (Figure 1d–f). Vehicle-treated SMC-MR-KO mice had significantly lower p-eNOS compared with vehicle-treated MR-intact mice (Figure 1d). Despite this, sorafenib treatment significantly decreased p-eNOS compared with vehicle in MR-intact mice as well as SMC-MR-KO mice (Figure 1d). Thus, SMC-MR appears to regulate the basal levels of p-eNOS but is not necessary for VEGFRi-induced resistance vessel EC dysfunction. Indeed, there was no significant difference in p-eNOS levels between sorafenib-treated SMC-MR-KO mice and MR-intact littermates. Sorafenib treatment induced significant glomerular endotheliosis compared with vehicle in both SMC-MR-KO and MR-intact mice, with no effect on genotype (Figure 1e). Once again, sorafenib treatment significantly increased SBP in both groups (subject: P < 0.0001; time: P < 0.0001), with no significant difference in SBP between the SMC-MR-Intact and SMC-MR-KO mice (Figure 1f). Thus, SMC-MR is not necessary for VEGFRi-induced resistance vessel EC dysfunction, renal impairment, or SBP elevation.

EC-MR contributes to VEGFRi-induced resistance vessel and renal dysfunction without impacting BP

Finally, EC-MR-KO mice were treated with sorafenib and resistance vessel eNOS activity, glomerular endotheliosis, and SBP were compared with MR-intact littermates (Figure 1h and i). Neither SMC-MR nor EC-MR genotype significantly modified ET-1 protein levels in mesenteric resistance vessels (Figures 1d and g, right). Once again, sorafenib treatment significantly decreased p-eNOS in resistance vessels of EC-MR-intact mice compared with vehicle-treated littermates yet, sorafenib did not significantly impact resistance vessel eNOS function in EC-MR-KO mice (Figure 1h). Additionally, EC-MR deletion did not affect basal eNOS activity in mice without VEGFRi treatment (Figure 1h). These data suggest that EC-MR-KO protects against sorafenib-induced EC dysfunction in mesenteric resistance arterioles. Renal glomerular endotheliosis was significantly increased in EC-MR-intact mice treated with sorafenib vs. vehicle (Figure 1i). In contrast, EC-MR-KO mice treated with sorafenib were protected from renal glomerular endotheliosis, as there was no difference in glomerular endotheliosis compared with littermate controls, resulting in significantly less glomerular endotheliosis in EC-MR-KO mice compared with MR-intact mice treated with sorafenib (Figure 1i). While SBP increased significantly in both groups (subject: P < 0.0001; time: P < 0.0001), there was no difference in SBP between sorafenib-treated EC-MR-Intact and -KO mice (Figure 1g). Sorafenib treatment increased SBP and diastolic BP with an associated decrease in heart rate. Spironolactone, SMC-MR-KO, and EC-MR-KO did not significantly impact the sorafenib-induced increase in BP or heart rate when averaged over 24 hours or when separated by light or dark phase of the diurnal cycle. These data collectively suggest that in the context of VEGFRi treatment, systemic inhibition of the MR and EC-specific MR KO were protective against renal impairment and attenuated resistance vessel eNOS dysfunction with no impact on BP.

DISCUSSION

This study investigated whether the MR plays a role in VEGFRi-induced hypertension using a mouse model in which sorafenib treatment significantly decreased resistance vessel p-eNOS activity, induced renal glomerular endotheliosis, and raised SBP, consistent with previous studies using other VEGFRis in mice.^8,15 Systemic MR inhibition with spironolactone decreased basal p-eNOS activity compared with placebo-treated mice, and this was reproduced by SMC-MR-specific KO, revealing a role for SMC-MR in maintaining basal resistance vessel eNOS activity. Moreover, spironolactone is protected from the sorafenib-induced decrease in p-eNOS and induction of glomerular endotheliosis. This vascular and renal protection was reproduced in EC-MR-KO mice. Finally, the acute hypertensive response to sorafenib was not impacted by MR blockade nor by specific KO of MR from SMCs or ECs. Thus, MR antagonism protects from components of VEGFRi-induced vascular and renal dysfunction, without impacting sorafenib-induced hypertension, and this protection is mediated specifically by MR in ECs.

VEGFRi treatment induces vascular and renal damage and hypertension in cancer patients and animal models,³ which was reproduced in this study using sorafenib in mice. This vascular and renal damage has been implicated as the cause of hypertension in VEGFRi-treated cancer patients. Here, decreased p-eNOS levels in mesenteric resistance vessels and increased renal glomerular endotheliosis were used as markers of vascular endothelial dysfunction and renal damage, respectively, as these markers have been shown to correlate with VEGFRi-induced hypertension in humans.^3,8,15 Here, MR antagonism protected mice from sorafenib-induced decreased eNOS activity and renal dysfunction, and this was recapitulated in EC-MR-KO mice, suggesting that EC-MR is necessary for sorafenib-induced endothelial dysfunction as measured by eNOS activity and glomerular endotheliosis. However, spironolactone and EC-MR-KO prevented vascular and renal dysfunction without attenuating the hypertensive response. There are certainly other components of vascular and renal function that were not tested in this study which could be contributing to the hypertensive response. However, these results suggest that reversing eNOS phosphorylation and glomerular endotheliosis is insufficient to decrease VEGFRi-induced hypertension in this model.

Ex vivo studies suggest that the EC-MR and SMC-MR may contribute to vascular function in opposing ways. For example, administering aldosterone directly into the lumen of a vessel to impact the endothelium resulted in vasorelaxation while adding it to the surrounding bath to target SMCs resulted in vasoconstriction.⁷ In that study, spironolactone treatment and SMC-MR-KO both lowered basal eNOS-S1177 phosphorylation. This finding adds an additional layer of complexity that includes intracellular crosstalk, as the presence of SMC-MR impacted eNOS activity which is generally thought to be expressed and to function in ECs. Indeed, in this study, SMC-MR-KO decreased basal eNOS phosphorylation (without VEGFRi) adding further support to the notion that MR in SMC contributes to basal EC eNOS activity.

Pharmacologic MR inhibition in renal proximal tubule cells and podocytes reduces proteinuria and promotes healthy function of the glomerular filtration barrier, demonstrating that MR plays a role in glomerular function.⁵ However, the role of MR in glomerular endotheliosis and renal injury has never been explored in the context of VEGFRi drug treatment. Preeclampsia is a hypertensive disorder of pregnancy that is induced by placental-derived factors that sequester VEGF.¹⁶ Biwer et al. recently demonstrated that SMC-MR contributes to enhanced sensitivity to hypertensive stimuli after exposure to preeclampsia induced by VEGF sequestration.¹⁰ However, in that model, SMC-MR did not impact the hypertensive response to VEGF deficiency during preeclampsia.¹⁰ This is consistent with our study in which SMC-MR-KO did not impact VEGFRi-induced hypertension. Rather, EC-MR, but not SMC-MR, prevented renal glomerular endotheliosis induced by sorafenib, suggesting that EC-MR contributes to sorafenib-induced renal injury.

Finally, MR is known to regulate vasomotor tone and BP through its effects on the vascular and renal systems.⁵ MR antagonists, including spironolactone, are effective at reducing BP, particularly when caused by increased aldosterone and in patients with resistant hypertension, but many of the benefits of therapeutic MR inhibition have been demonstrated to be independent of a change in BP.⁵ While the renin–angiotensin–aldosterone system remains understudied in the context of VEGFRi therapy, several small clinical studies reveal that VEGFRi-mediated hypertension is associated with a decrease in renin activity without a change in aldosterone levels, raising the potential for dysregulated aldosterone production.¹⁷ In addition, angiotensin-II and reactive oxygen species have been shown to increase with VEGFRi treatment and may contribute to the pathogenesis of vascular and renal toxicity of these therapies.^15,18 Overall, the role of the renin–angiotensin–aldosterone system in VEGFRi-induced hypertension warrants further investigation.

All the mice in our study experienced a significant and sustained increase in SBP with sorafenib treatment which was not significantly impacted by spironolactone treatment, EC-MR, or SMC-MR knockout status. Taken together, these data suggest that while EC-MR is necessary for VEGFRi-induced impairment in eNOS phosphorylation and renal glomerular endotheliosis in mice, neither impaired eNOS activity nor renal glomerular endotheliosis is necessary for VEGFRi-induced hypertension.

This study has several limitations. Since VEGFRi-induced hypertension occurs on a short timescale (days) in humans and mice,³ we did not model the longer-term effects of EC-MR-KO, SMC-MR-KO, and spironolactone treatment. It is possible that the effects of pharmacologic MR blockade or MR genetic deletion on VEGFRi-induced BP alterations require a longer time period to become evident and future studies will be needed to clarify this.¹⁰ In addition, while vehicle controls were not included here, we previously confirmed that the vehicle used to administer sorafenib does not alter BP, glomerular endotheliosis, or resistance vessel protein changes associated with sorafenib treatment.¹⁹ While it is established that eNOS activity is associated with vascular endothelial function and vasomotor tone,⁷ we did not specifically measure endothelium-dependent relaxation to acetylcholine nor SMC-mediated myogenic tone in the mesenteric resistance vessels of these mice. Future studies are needed to clarify the impact of VEGFRis on resistance vasomotor function and how EC-MR and SMC-MR may contribute. Glomerular endotheliosis is a histologic characteristic of renal dysfunction that is associated with proteinuria and impaired natriuresis in human patients with VEGFRi-induced hypertension.^8,16 In this study, we did not measure changes in 24-hour urine protein nor pressure natriuresis integrity, factors that could be assessed in future studies as additional measures of kidney function. To confirm the translational relevance of these findings, further studies in larger animal models that more closely resemble human BP physiology are also warranted.

In summary, using a mouse model of VEGFRi-induced hypertension with impaired eNOS activity in resistance vessels and increased renal glomerular endotheliosis, this study demonstrates for the first time, a role for MR specifically in ECs in VEGFRi-induced vascular and renal damage. MR does not appear to be necessary for VEGFRi-induced hypertension. As cancer treatment improves and patients live longer, cardiovascular disease has become the leading cause of death in cancer survivors.²⁰ Preventing vascular and kidney damage caused by cancer therapies may have benefits for cardiovascular health that would improve outcomes in cancer survivors. The findings that spironolactone is renal protective and stabilizes eNOS activity in the context of VEGFRi treatment supports further testing of MR inhibition as a renal and vascular protective agent alongside effective antihypertensive therapy in the growing populations of cancer patients for which VEGFRis are indicated.

FUNDING

This work was supported by R01 CA243542 (NCI), HL119290 (NHLBI), and HL095590 (NHLBI) to I.Z.J.

CONFLICT OF INTEREST

I.Z.J is a consultant for Boehringer Ingelheim. All other authors have no disclosures.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.