Oxidative Stress Contributes to Microvascular Endothelial Dysfunction in Men and Women with Major Depressive Disorder

Abstract

Rationale:

In rodent models of depression, oxidative stress-induced reductions in nitric oxide (NO) bioavailability contribute to impaired endothelium-dependent dilation. Endothelial dysfunction is evident in Major Depressive Disorder (MDD); however, the molecular mediators remain undefined.

Objective:

We sought to translate preclinical findings to humans by testing the role of oxidative stress in mediating microvascular endothelial dysfunction, including potential modulatory influences of sex, in MDD.

Methods and Results:

Twenty-four treatment-naïve, otherwise healthy, young adults with MDD (14 women; 18–23 yrs) and twenty healthy adults (HC; 10 women; 19–30 yrs) participated. Red blood cell flux (laser Doppler flowmetry) was measured during graded intradermal microdialysis perfusion of the endothelium-dependent agonist acetylcholine, alone and in combination with an NO synthase inhibitor (L-NAME), a superoxide scavenger (Tempol), and an NADPH oxidase inhibitor (apocynin), as well as during perfusion of the endothelium-independent agonist sodium nitroprusside. Tissue oxidative stress markers (e.g., nitrotyrosine abundance, superoxide production) were also quantified. Endothelium-dependent dilation was blunted in MDD and mediated by reductions in NO-dependent dilation. Endothelium-independent dilation was likewise attenuated in MDD. In MDD, there were no sex differences in either NO-mediated endothelium-dependent dilation or endothelium-independent dilation. Acute scavenging of superoxide or inhibition of NADPH oxidase improved NO-dependent dilation in MDD. Expression and activity of oxidative stress markers were increased in MDD. In a subset of adults with MDD treated with a selective serotonin reuptake inhibitor for their depressive symptoms and in remission (n=8; 7 women; 19–37 yrs), NO-mediated endothelium-dependent dilation was preserved, but endothelium-independent dilation was impaired, compared to HC.

Conclusions:

Oxidative stress-induced reductions in NO-dependent dilation, as well as alterations in vascular smooth muscle function, directly contribute to microvascular dysfunction in MDD. Strategies targeting vascular oxidative stress may be viable therapeutic options for improving NO-mediated endothelial function and reducing cardiovascular risk in MDD.

INTRODUCTION

Depression, defined as either Major Depressive Disorder (MDD) or significant depressive symptoms with substantial functional impairment¹, occurs in ~10–15% of the population² and is projected to become the second leading cause of worldwide disease and disability by 2030³. Compelling evidence now links depressive disorders to the excessive (~3x greater risk) and premature (~7.5 yrs earlier) development of cardiovascular disease (CVD), independent of traditional risk factors^4–6. Interestingly, CVD risk remains elevated even with the remission of depressive symptoms⁷. Thus, otherwise healthy adults with depression represent a clinically important at-risk cohort for mechanism-specific therapeutic interventions to prevent or mitigate the development of CVD.

Although multiple factors undoubtedly contribute, extensive evidence indicates that vascular dysfunction plays a pathogenic role in depression-CVD comorbidity. MDD is associated with impairments in vascular function graded in relation to the severity of depressive symptoms^8–11. Depression-induced impairments in brachial artery endothelium-dependent dilation [flow-mediated dilation (FMD)] are indicative of reductions in vascular nitric oxide (NO) bioavailability and function^{8, 11–14}. These functional vascular measures are consistent with the generalized reductions in plasma NO metabolite concentration reported in MDD^{15, 16}. However, the specific mechanisms underlying reductions in NO-mediated endothelial function in MDD remain relatively undefined. In rodent models of depression^17–19, the deficit in endothelial function is due, in part, to increased vascular oxidative stress—specifically superoxide—and subsequent attenuation of NO-dependent dilation. In depressed adults, biomarkers of oxidative stress (e.g., plasma lipid peroxidation, platelet protein carbonylation) are increased^{20, 21}. However, no studies have directly examined NO-mediated vascular function in vivo in MDD using targeted pharmacological approaches, nor have any studies investigated the potential sources of, or a role for, oxidative stress in mediating endothelial dysfunction in otherwise healthy adults with depression.

The link between depression and CVD may be modulated by sex. Compared to age-matched men, premenopausal women have lower CVD risk²², yet are more than twice as likely to suffer from depressive disorders²³. However, epidemiological data also suggest that MDD is associated with a greater CVD incidence in women compared to men^{6, 24}. These apparently contradictory notions implicate potential sex differences in the pathophysiological mechanisms underlying MDD-CVD comorbidity. In this regard, female rodents display increased behavioral susceptibility to depression, but less severe impairments in endothelial function compared to their male counterparts^{17, 19}, owing to greater antioxidant defense mechanisms^{17, 19, 25}. Despite this compelling pre-clinical evidence, to our knowledge no investigations have examined potential sex differences in the mechanistic underpinnings of endothelial dysfunction in human MDD. Such studies are critical in order to develop sex-specific preventative and therapeutic interventions.

The aim of the present investigation was to directly examine the mechanisms mediating microvascular endothelial dysfunction, including the modulatory influence of sex, in treatment-naïve, otherwise healthy, young adults with MDD compared to healthy non-depressed men and women (HC). We systematically examined endothelium-dependent dilation, vascular smooth muscle sensitivity, and the effect of acute pharmacological inhibition of oxidative stress on endothelial function in the cutaneous microcirculation—a validated bioassay for systemic microvascular function^26–30. Complementary to our direct pharmacological assessment of the molecular mediators of microvascular function, we performed ex vivo analyses of endothelium-dependent signaling pathways and oxidative stress-related enzyme abundance/expression and activity in cutaneous tissue homogenates obtained from MDD and HC adults. We hypothesized that MDD adults would have (1) attenuated endothelium-dependent dilation, (2) reduced NO-mediated dilation, and (3) increased vascular oxidative stress. Given the aforementioned preclinical data indicating a role for superoxide in mediating endothelial dysfunction in depression, we further hypothesized that acute localized antioxidant treatment with a scavenger of superoxide or inhibition of NADPH oxidase-derived superoxide production would improve endothelium-dependent dilation in MDD via increased vascular NO bioavailability. We also hypothesized that endothelial dysfunction would be less severe in MDD women compared to men, due to greater antioxidant defense mechanisms and a relative preservation of NO function. Finally, because treatment with selective serotonin reuptake inhibitors (SSRI) for the management of depressive symptoms may have pleiotropic vasculo-protective effects^31–33 and in an effort to gain preliminary insight to the relation between depressive symptoms and microvascular dysfunction, we also examined endothelium-dependent and –independent dilation in a subset of adults with MDD treated with a SSRI for their depression and in remission.

METHODS

A detailed Methods and Materials section is available in the Online Data Supplement. All experimental procedures and protocols were approved by The Institutional Review Board at The Pennsylvania State University. The investigation was conducted in accordance with the Declaration of Helsinki. The nature, risks, and benefits of all study procedures were explained to volunteers, and their verbal and written informed consent was obtained voluntarily prior to participation.

A total of 44 participants were tested: 20 healthy adults (10 men) without any history of major psychiatric illness and 24 otherwise healthy MDD patients (10 men) with a current symptomatic major depressive episode (non-medicated; Online Table I). Additionally, 8 adults diagnosed with MDD and currently treated with a SSRI for their depressive symptoms (monotherapy) were also tested (Online Table II). All SSRI-treated adults were treated for at least 8 weeks prior to study enrollment, and their MDD was in remission, confirmed by the absence of clinically meaningful depressive symptoms.

Assessment of microvascular function (in vivo).

Using sterile technique, intradermal microdialysis probes (CMA Linear 30 probe, 6 kDa; Harvard Apparatus, Holliston, MA, USA) were inserted into the dermal layers of the ventral forearm for the local delivery of pharmacological agents, as previously described³⁴. Red blood cell flux, an index of cutaneous blood flow, was continuously measured directly over each microdialysis site with an integrated laser Doppler flowmeter probe placed in a local heating unit (VP12 and VHP2; Moor Instruments, Wilmington, DE, USA) set to thermoneutrality (33°C), unless noted otherwise. Automated brachial blood pressure (Cardiocap; GE Healthcare, Milwaukee, WI, USA) was measured every 5 min.

Endothelium-dependent dilation was assessed by perfusion of ascending concentrations of acetylcholine (ACh; 10⁻¹⁰ – 10⁻¹ mol/L; United States Pharmacopeia, Rockville, MD, USA) alone and during concurrent administration of 15 mmol/L N^G-nitro-L-arginine methyl ester (L-NAME; Calbiochem, EMD Millipore, Billerica, MA, USA) to non-selectively inhibit NO synthase (NOS), 10 μmol/L 4-Hydroxy-Tempo (Tempol; Sigma-Aldrich, St. Louis, MO, USA) to scavenge superoxide, 100 μmol/L apocynin (Bio-Techne Corporation, Minneapolis, MN, USA) to inhibit NADPH oxidase, Tempol+L-NAME, or apocynin+L-NAME^34–38. Endothelium-independent dilation was assessed by perfusion of ascending concentrations of SNP (10⁻⁷ – 10⁻¹ mol/L; United States Pharmacopeia). Following each dose-response protocol, maximal dilation was elicited^{34, 39}. Details are provided in the Online Supplement.

Assessment of expression and activity of NO and oxidative stress (in vitro).

In a subset of participants (n=8 HC, n=8 MDD), two 3 mm cutaneous vascular tissue samples were obtained from the ventral forearm under local anesthesia (2% lidocaine without epinephrine) via punch biopsy, immediately snap frozen in liquid nitrogen, and stored at −80°C until analysis^{34, 40}. Protein abundance/expression was determined by Western blot and total reactive oxygen species and reactive nitrogen species (ROS+RNS) production was quantified fluorometrically (Cell Biolabs Inc, San Diego, CA, USA). The relative contribution of superoxide to total ROS+RNS was assessed by pretreatment with the SOD mimetic MnTBAP (Sigma-Aldrich). Details are provided in the Online Data Supplement.

Data and statistical analysis.

Data were recorded at 40 Hz (PowerLab and LabChart; ADInstruments, Bella Vista, NSW, Australia). Vascular conductance was calculated as laser Doppler flux (perfusion units) divided by mean arterial pressure, normalized as a percentage of the site-specific maximum (%CVC_max), and averaged during 5 min of baseline and during the last min of each ACh or SNP dose. The NO-mediated component of ACh-induced dilation was calculated as the difference between control/Tempol/apocynin and the corresponding L-NAME-treated site (arbitrary units)^{34, 40}.

Student’s unpaired t-tests were used to compare subject characteristics and biochemical data. Sigmoidal dose-response curves with variable slope were generated using non-linear regression modelling (Prism; GraphPad, San Diego, CA, USA). Differences between groups and microdialysis sites in the logEC₅₀ (effective concentration resulting in 50% of the maximal response) were analyzed using an F-test for repeated-measures comparisons^{40, 41}. NO-dependent dilation was analyzed using two-way (group x. pharmacological agent) mixed model repeated-measures ANOVA, with post hoc Bonferroni corrections applied for specific planned comparisons (SAS v9.4; Cary, NC, USA). If necessary, these statistical analyses were also performed separately with the inclusion of the SSRI-treated group. Values are presented as mean±SEM and significance was set at α<0.05.

RESULTS

Participant characteristics.

Participants in the HC and MDD groups were well-matched for anthropometric characteristics, habitual physical activity, resting hemodynamics, and blood biochemistry (all p>0.05), though MDD were significantly younger than HC (Online Table I; p=0.01). Adults with MDD were experiencing a major depressive episode of mild-moderate severity (Online Table I; p<0.01).

NO-mediated endothelium-dependent and -independent dilation is blunted in MDD.

Neither baseline nor SNP-induced maximal vascular conductance was different between groups or pharmacological treatments (Online Table III; all p>0.05). Endothelium-dependent dilation was blunted in MDD (Fig. 1; −6.4±0.3 HC vs. −4.9±0.5 logEC₅₀ MDD; p<0.01) and negatively related to depressive symptom severity (R²=0.28; p<0.001). NOS inhibition blunted ACh-induced dilation in HC (−6.4±0.3 control vs. −4.6±0.7 logEC₅₀ L-NAME; p=0.013) but had no effect in MDD (Fig. 1; −4.9±0.5 control vs. −4.7±0.6 logEC₅₀ L-NAME; p=0.749), indicating significant reductions in the functional contribution of NO to ACh-induced vasodilation in MDD (p<0.001). VSMC sensitivity to exogenous NO (endothelium-independent dilation) was also reduced in MDD (Fig. 2; −6.3±0.2 HC vs. −5.0±0.2 logEC₅₀ MDD; p<0.01). There were no group differences in expression of eNOS, phosphorylated eNOS, VASP, or pVASP (Online Fig. I, II; all p>0.05).

Figure 1.

Vascular conductance in response to increasing doses of acetylcholine (ACh) alone (CON; white symbols) and during concurrent nitric oxide (NO) synthase inhibition (L-NAME, N^G-nitro-L-arginine methyl ester; black symbols) in healthy adults (HC; Panel A) and in those with Major Depressive Disorder (MDD; Panel B). The NO-dependent component of ACh-induced dilation is indicated by gray shading. Endothelium-dependent dilation was blunted in MDD and mediated by reductions in NO-mediated dilation. Data are mean ± SE. *p<0.05 vs. HC; †p<0.05 vs. CON.

Figure 2.

Vascular conductance in response to increasing doses of sodium nitroprusside (SNP) in healthy adults (HC; white symbols) and in those with Major Depressive Disorder (MDD; black symbols). Endothelium-independent dilation was attenuated in MDD. Data are mean ± SE. *p<0.05 vs. HC.

Sex does not modulate NO-dependent microvascular endothelial function in MDD.

There were no differences in ACh-induced vasodilation between HC men and women (Online Fig. IIIA; −6.3±0.4 HC men vs. −6.2±0.5 logEC₅₀ HC women; p>0.05). Endothelium-dependent dilation was attenuated to a greater extent during concurrent NOS inhibition in HC men (Online Fig. IIIA; −3.7±1.0 HC men vs. −6.1±0.8 logEC₅₀ HC women; p<0.01), reflecting modestly greater NO-dependent dilation. In HC, there were no sex differences in endothelium-independent dilation (Online Fig. IVA; −6.3±0.3 HC men vs. −6.3±0.2 logEC₅₀ HC women; p=0.92).

Depressive symptom severity was not different between MDD men and women (PHQ-9; 10±2 men vs. 13±2 women; p=0.42). In MDD, there were no sex differences during perfusion of ACh alone (−4.9±0.9 MDD men vs. −4.9±0.5 logEC₅₀ MDD women; p=0.97) or during concurrent NOS inhibition (Online Fig. IIIB; −5.1±0.7 MDD men vs. −4.4±0.7 logEC₅₀ MDD women; p=0.51), suggesting that sex does not modulate the NO contribution to ACh-induced dilation in MDD (p=0.96). Therefore, further analyses of potential sex differences in the role of oxidative stress in contributing to impaired NO-dependent dilation were not performed. Further, there were no sex differences in VSMC sensitivity to exogenous NO (Online Fig. IVB; −4.9±0.4 MDD men vs. −5.1±0.2 logEC₅₀ MDD women; p=0.82).

Oxidative stress mediates attenuated NO-dependent endothelial function in MDD.

As expected, because increased vascular oxidative stress is not evident in healthy young adults, perfusion of Tempol (Fig. 3A, C) or apocynin (Fig. 3D, F) had no effect on ACh-induced vasodilation in HC (−6.4±0.3 control; −4.5±1.3 Tempol; −4.8±0.3 apocynin logEC₅₀; all p>0.05). In MDD, the effect of acute treatment with Tempol (−4.9±0.5 control vs. −8.6±3.5 Tempol logEC₅₀; p=0.75) or apocynin (−4.9±0.5 control vs. −5.4±0.3 apocynin logEC₅₀; p=0.32) on microvascular sensitivity to ACh were not statistically significant (Fig. 3B, E). However, there were also no differences in endothelium-dependent dilation during NOS inhibition in the presence or absence of either Tempol (Fig. 3B; −4.7±0.6 L-NAME vs. −4.1±0.4 Tempol+L-NAME logEC₅₀; p=0.08) or apocynin (Fig. 3E; −4.7±0.6 L-NAME vs. −3.9±0.4 apocynin+L-NAME logEC₅₀; p=0.30). Thus, scavenging superoxide or inhibiting NADPH oxidase significantly improved the NO-dependent component of ACh-induced dilation in MDD (Fig. 3C, F; both p<0.01), indicating that pharmacological suppression of oxidative stress acutely improved NO function within the microvasculature.

Figure 3.

Vascular conductance in response to increasing doses of acetylcholine (ACh) alone (CON) and during concurrent perfusion of the nitric oxide (NO) synthase inhibitor N^G-nitro-L-arginine methyl ester (L-NAME) and the antioxidants Tempol (TEM) and apocynin (APO), both alone and in combination with L-NAME, in healthy adults (HC; Panels A, D) and in those with Major Depressive Disorder (MDD; Panels B, E). The NO-dependent component of ACh-induced dilation, expressed as the difference in area under the curve between CON/TEM/APO and the corresponding L-NAME site, is presented in Panels C and F. In MDD, acute treatment with the TEM or APO improved NO-dependent dilation. Data are mean ± SE. *p<0.05 vs. HC; †p<0.05 vs. CON; ǂp<0.05 vs. L-NAME-treated site.

Plasma oxidized LDL concentration was not different between groups (26.1±4.0 HC vs. 19.5±1.8 U/L MDD; p=0.15). There were no group differences in protein expression of either SOD1 (p=0.61) or SOD2 (p=0.84) in cutaneous tissue homogenates (Online Fig. V). However, nitrotyrosine abundance (p=0.01) and expression of the p47^phox subunit of NADPH oxidase (p=0.02) were increased in MDD (Fig. 4, 5). Further, total ROS+RNS production (p<0.01), as well as superoxide activity (p=0.03), were substantially increased in MDD (Fig. 6), consistent with the notion that increased vascular oxidative stress mechanistically contributes to deficits in NO function and concomitant reductions in endothelium-dependent dilation in MDD.

Figure 4.

Representative Western blots (Panel A) and group summary data (Panel B) for nitrotyrosine (NT) abundance in cutaneous tissue homogenates of healthy adults (HC; white bars) and those with Major Depressive Disorder (MDD; black bars). NT abundance was increased in MDD. GAPDH, glyceraldehyde 3-phosephate dehydrogenase. Data are mean ± SE. *p<0.05 vs. HC.

Figure 5.

Representative Western blots (Panel A) and group summary data (Panel B) for protein expression of the p47^phox subunit of NADPH oxidase in cutaneous tissue homogenates of healthy adults (HC; white bars) and those with Major Depressive Disorder (MDD; black bars). Expression of p47^phox was increased in MDD. GAPDH, glyceraldehyde 3-phosephate dehydrogenase. Data are mean ± SE. *p<0.05 vs. HC.

Figure 6.

Total reactive oxygen and nitrogen species (ROS+RNS; Panel A) and superoxide production (Panel B) in cutaneous tissue homogenates of healthy adults (HC; white bars) and those with Major Depressive Disorder (MDD; black bars). Increased superoxide production was evident in MDD. GAPDH, glyceraldehyde 3-phosephate dehydrogenase. Data are mean ± SE. *p<0.05 vs. HC.

NO-mediated endothelium-dependent dilation is not impaired in SSRI-treated adults with MDD.

SSRI-treated adults were not currently suffering from a major depressive episode (Online Table II). Endothelium-dependent dilation was not different between SSRI-treated and HC adults during perfusion of ACh alone (Fig. 7; −6.4±0.3 HC vs. −6.4±0.2 SSRI logEC₅₀; p=0.90) or during concurrent NOS inhibition (−4.7±0.7 HC vs. −3.8±2.2 SSRI logEC₅₀; p=0.92), indicating functionally intact NO-dependent dilation in SSRI-treated adults with MDD in remission. However, SNP-induced dilation was reduced in SSRI-treated adults (Fig. 8; −6.3±0.2 HC vs. −5.1±0.2 logEC₅₀ SSRI; p<0.01).

Figure 7.

Vascular conductance in response to increasing doses acetylcholine (ACh) alone (CON; Panel A) and during concurrent nitric oxide (NO) synthase inhibition (L-NAME, L-N^G-nitroarginine methyl ester; Panel B) in a subset of adults with Major Depressive Disorder who were treated with a selective serotonin reuptake inhibitor and in remission (MDD+SSRI; n=8). ACh-induced dilation in healthy adults (HC) and untreated adults with MDD is presented for reference. Endothelium-dependent dilation in SSRI-treated adults was not different from that in HC, due to a relative preservation of NO-mediated dilation.

Figure 8.

Vascular conductance in response to increasing doses of sodium nitroprusside (SNP) in a subset of adults with Major Depressive Disorder who were treated with a selective serotonin reuptake inhibitor and in remission (MDD+SSRI; n=8). SNP-induced dilation in healthy adults (HC) and untreated adults with MDD is presented for reference. Endothelium-independent dilation was attenuated in SSRI-treated adults. Data are mean ± SE. *p<0.05 vs. HC.

DISCUSSION

This is the first study in humans to directly assess the contribution of superoxide to impaired NO-mediated microvascular function in MDD using a targeted in vivo pharmacological approach. The present findings demonstrate significantly attenuated endothelium-dependent dilation in MDD adults with mild-moderate symptomology that is mediated, in part, by reductions in NO-dependent mechanisms. Endothelium-independent dilation was also blunted in MDD, suggesting that alterations in VSMC sensitivity to NO also contribute to microvascular dysfunction. However, contrary to our initial hypothesis, there were no sex differences in either NO-mediated endothelium-dependent dilation or endothelium-independent dilation in MDD. Further, in MDD, acute administration of either a superoxide scavenger or a NADPH oxidase inhibitor improved NO-dependent dilation. Taken together with the ex vivo biochemical analyses demonstrating increased abundance/expression and activity of enzymes associated with an increased oxidative load, these data support the hypothesis that increased vascular oxidative stress contributes to microvascular dysfunction in MDD. Finally, NO-mediated endothelium-dependent dilation was preserved, but endothelium-independent dilation was impaired, in SSRI-treated adults with MDD in remission. Collectively, these findings provide support for the concept that ROS-induced reductions in NO-dependent mechanisms directly contribute to microvascular dysfunction in MDD and further suggest that targeting oxidative stress — specifically superoxide —may be a viable therapeutic strategy for improving vascular function and reducing CVD risk in adults with depression.

In rodent models, the development of a depressive phenotype is causally linked to impairments in endothelium-dependent dilation in both the conduit and resistance vasculature^{17–19, 42–44}. These functional deficits are specific to the endothelium, as dilation in response to exogenous NO and adrenergic vasoconstriction are not different from that in healthy control rodents^{18, 19, 42, 44}. Mechanistically, blunted endothelium-dependent dilation is mediated in large part by impairments in NO bioavailability and function^{17–19, 42–44}. Extensive preclinical evidence further suggests that excessive superoxide production contributes to depression-induced endothelial dysfunction^{17–19, 45, 46}. Further, treatment of aortic segments isolated from depressed rodents with a superoxide scavenger improves endothelium-dependent dilation via increases in vascular NO production^{17, 19, 45}. These data suggest that increased ROS generation, specifically superoxide, directly limits NO bioavailability and mechanistically contributes to endothelial dysfunction in rodent models of depression.

Few studies have examined vascular function in otherwise healthy young adults with MDD and free of any other cardiovascular risk factors. Of these, the majority have reported depression-induced deficits in gross measures of endothelial function^{8, 11–14}, though this is not a universal finding¹⁶. Most studies have assessed endothelial function utilizing a reactive hyperemia-induced increase in endothelial shear stress, typically via brachial artery FMD. Although NO mediates ~50% of conduit artery FMD in young healthy individuals⁴⁷, making it a valid measure of endothelium-dependent vasomotor function in the human macrovasculature, the technique does not allow for the direct quantification of NO-dependent dilation nor does it permit investigation of the specific mechanisms mediating endothelial dysfunction. We utilized the human cutaneous circulation to probe the mechanisms mediating microvascular dysfunction in MDD. The cutaneous circulation is a validated model for generalized microvascular function and has been utilized to examine the mechanisms underlying microvascular dysfunction in multiple pathologies^26–30. Moreover, disease-related alterations in microvascular function often first appear in the cutaneous circulation, making it an accessible circulatory bed in which to examine mechanistic alterations in vascular function as a means for the early identification of patients at risk for the future development of CVD, before overt target-organ damage becomes apparent²⁹.

Using this model, and consistent with reports of large artery dysfunction in depression, the present findings demonstrate marked impairments in microvascular ACh-induced dilation in MDD. Non-selective pharmacological inhibition of NOS attenuated ACh-induced dilation in HC, but not in MDD, indicating that a functional absence of NO-mediated dilation contributes to endothelial dysfunction in MDD. However, to further examine potential MDD-induced reductions in NO bioavailability, we measured expression of eNOS and phosphorylated eNOS in cutaneous tissue homogenates. We did not detect a reduction in phosphorylated eNOS in MDD. Although perhaps surprising, these ex vivo data are consistent with reports of unaltered eNOS expression in rodent models of depression¹⁸. Therefore, although the use of a non-selective NOS inhibitor precludes conclusions regarding the isoform-specific source of NO in the vasculature, w nevertheless interpret these data as suggesting that vascular NO production per se is not reduced in MDD. Instead, it appears likely that mechanisms secondary to NO production and diffusion modulate endothelial dysfunction in MDD.

In this regard, excessive superoxide rapidly oxides and inactivates NO⁴⁸. Thus, building on preclinical data^{17–19, 45, 46}, we sought to elucidate whether increased vascular superoxide directly contributes to blunted NO-dependent dilation in human MDD. Acute localized treatment with Tempol, a scavenger of superoxide anion⁴⁹, improved NO-mediated dilation in MDD. These data suggest that vascular superoxide mechanistically contributes to endothelial dysfunction in MDD. However, Tempol is not specific for superoxide and has been shown to metabolize other ROS (e.g., hydrogen peroxide, hydroxyl radicals, and peroxynitrite)³⁷. To circumvent this issue, we also assessed endothelial function following acute administration of apocynin to inhibit NADPH oxidase³⁸, a major source of superoxide in the endothelium⁵⁰. Similar to Tempol, treatment with apocynin improved NO-mediated endothelium-dependent dilation. Interestingly, the improvements in endothelial function following acute ‘antioxidant’ treatment were only evident for NO-mediated dilation, as neither Tempol nor apocynin alone directly improved endothelial sensitivity to ACh. The reason(s) underlying this seemingly paradoxical finding is not immediately clear. Nevertheless, these data suggest that increased vascular ROS production, specifically NADPH oxidase-derived superoxide, directly contributes to impairments in NO-mediated signaling mechanisms and subsequent endothelial dysfunction in human depression.

Consistent with our functional data indicating a role for excessive microvascular superoxide production, increased circulating markers of oxidative stress and decreased antioxidant markers have been reported in MDD^{20, 21, 51, 52}. We did not detect MDD-related elevations in plasma oxidized LDL, a generalized biomarker of systemic oxidative damage to lipids⁵³. Because ROS have a short half-life, there is inherent difficulty in making gross measurements and these are not necessarily reflective of tissue concentrations. In addition, there is a wide range of oxidative stress biomarkers, the type of biological specimen used, and laboratory techniques for measurement, each with its own strengths and limitations. However, to complement the measure of whole-body redox control, we also measured tissue-specific expression of oxidant and antioxidant enzymes, as well as total ROS+RNS production. Consistent with previous findings²¹, nitrotyrosine abundance and expression of the p47^phox subunit of NADPH oxidase were increased in cutaneous tissue homogenates from MDD. Further, total ROS+RNS activity and superoxide generation specifically were markedly increased in MDD. We additionally expected to demonstrate a reduction in SOD expression in MDD. There were no group differences in expression of the antioxidant enzymes SOD1 and SOD2. These data, however, do not rule out potential deficits in SOD activity and/or post-translational modifications. Regardless, it appears that superoxide production in MDD outstrips the ability of SOD to handle the oxidative load, especially in the absence of a compensatory upregulation in antioxidant defense mechanisms. We acknowledge that the cutaneous tissue homogenates represent mixed tissue (e.g., blood vessels, keratinocytes, nerve endings, etc.) and provide limited ability to specifically assess protein expression/activity in endothelial or VSMC alone. Nevertheless, when considered collectively with the functional in vivo data, the present findings build on and extend those from preclinical models and strongly suggest that increased NADPH oxidase-derived superoxide directly contributes to MDD-induced microvascular endothelial dysfunction via reductions in NO-mediated mechanisms.

Unexpectedly, expression of pVASP, which is reflective of downstream eNOS-cGMP signaling in the VSMC⁵⁴, was not reduced in cutaneous tissue homogenates from MDD patients. This finding may appear incongruous with the functional data, as it suggests that the primary mechanism contributing to reductions in NO-dependent dilation in MDD was not eNOS uncoupling. However, there are other downstream mechanisms by which superoxide may influence VSMC function. As such, our interpretation of these data is that superoxide is likely exerting both NO-dependent, and NO-independent, effects. Indeed, vascular oxidative stress has been linked to impairments in cGMP-dependent protein kinase activity in VSMCs themselves, leading to attenuated relaxation⁵⁴. Thus, it is feasible that the persistent and elevated superoxide generation in MDD directly impairs dilator mechanisms and/or concurrently activates myogenic constrictor pathways within VSMCs⁵⁵, independent of NO, thus further propagating the oxidative burden on the microvasculature.

This study did not probe the causative link between MDD and increased vascular oxidative stress, which is a limitation. Relatively few studies have examined the specific mechanism(s) by which MDD results in increased NADPH oxidase expression and activity. Angiotensin II is a potent stimulator of NADPH oxidase⁴⁸ and, importantly, dysregulation of angiotensin II signaling has been described in depression⁵⁶. Treatment with an angiotensin II receptor antagonist reduces oxidative stress and exerts antidepressant-like effects in rodent models⁵⁷. In humans, most studies suggest that drugs targeting the renin-angiotensin-aldosterone system improve depressive symptoms⁵⁸. MDD-induced alterations in the renin-angiotensin-aldosterone system may trigger excessive activation of vascular NADPH oxidase having deleterious consequences for endothelial function. Further, there is strong and broad support for the general hypothesis that inflammation plays a critical role in the pathogenesis of MDD, at least in some individuals⁵⁹. Inflammatory cytokines, in turn, regulate vascular NADPH oxidases, linking inflammation to oxidative stress. The potential for angiotensin II and/or inflammatory cytokines to contribute an upregulation of NADPH oxidase in MDD remains to be examined.

Our data also demonstrate significant impairments in SNP-induced endothelium-independent dilation in MDD, suggesting a loss of sensitivity to NO in the VSMC itself also contributes to the marked microvascular dysfunction in MDD. This finding was unexpected given the preservation of 1) SNP-induced dilation documented in rodent models^{18, 19} and 2) nitroglycerin-induced brachial artery dilation in adults with MDD^{8, 11}. However, ROS, particularly NADPH oxidase-derived superoxide, play a major role in VSMC tone, particularly in the microcirculation⁶⁰. Thus, given that alterations in microvascular function are evident prior to impairments in large conduit arteries, it is conceivable that our approach allowed us to detect early MDD-related impairments in the ability of the VSMC to respond to NO before similar functional deficits can be observed in other vascular beds. However, in the present study design, it is not possible to determine the extent to which impairments in endothelial or VSMC function in isolation—or in combination—contribute to the microvascular dysfunction evident in MDD.

Young women are more than twice as likely to develop MDD as men and also suffer greater depressive symptom severity²³. However, premenopausal women have lower CVD risk than men, despite the presence of comparable risk factors, including depression²², thus implicating sex differences in the pathophysiology underlying the MDD-CVD link. In preclinical models, investigators have demonstrated a greater susceptibility to depressive symptoms but less severe impairments in endothelial function in female compared to male rodents^{17, 19, 45}. This relative preservation of endothelial function in female rodents has been attributed to greater antioxidant defense mechanisms and subsequent maintenance of NO bioavailability and function¹⁹. However, to date, no studies have examined sex differences in the mechanisms underlying vascular dysfunction in otherwise healthy young adults with MDD.

In contrast to our hypothesis, there were no sex differences in ACh-induced dilation alone or during concurrent NOS inhibition in adults with MDD. Further, there were no sex differences in SNP-induced dilation. These data suggest that sex does not substantially modulate the relative contribution of NO to vasodilatory mechanisms in the microvasculature in MDD. It is important to note that the young women in the present study were studied without regard for menstrual cycle phase or oral contraceptive use. Recent studies suggest that endothelial function is stable across the menstrual and hormonal contraceptive cycles^{61, 62}. Changes in female sex hormones do not appear to influence ACh-induced dilation in the cutaneous microcirculation^{62, 63}. In addition, rodent models of depression demonstrate that endothelial dysfunction is apparent throughout the estrous cycle, despite a varied hormonal profile¹⁹. Because a detailed examination of the influence of cyclic variation in estrogen on endothelial function was not the primary purpose of this study, we instead probed for a “true” sex difference in NO-mediated vascular function. Further, because MDD adults were non-medicated, experimental visits were scheduled immediately following enrollment, and investigators encouraged all MDD patients to expediently seek clinical follow-up.

We acknowledge that the MDD adults in the present study presented a limited range of depressive symptom severity and recognize that this sample reflects only part of the spectrum of MDD and therefore may not be representative of more severely ill patients or patients with additional psychiatric and medical comorbidities. Nevertheless, we detected marked impairments in NO-mediated microvascular function in this MDD cohort, which are comparable in magnitude to the deficits reported in adults with traditional CVD risk factors^{34, 64, 65}. Future studies are required to determine the potential relation between depressive symptom severity and the mechanistic mediators of endothelial dysfunction in MDD.

In this study, microvascular function was assessed in adults with MDD during a current depressive episode, limiting our ability to make inferences regarding the temporal relation between depressive symptom development and microvascular dysfunction. In a recent meta-analysis¹³, the majority of included studies reported a significant inverse relation between depression and endothelial function, regardless of whether of depression was current or remitted. Although a prospective longitudinal study conducted over the course of several years in adults with MDD could better addressed the temporal relation between symptomology and microvascular dysfunction, the inability to predict depressive episode onset and remission makes the logistics of these types of studies challenging. In an effort to begin to address this question, we assessed endothelium-dependent and –independent dilation in a cohort of adults with MDD currently treated with a SSRI to manage their depressive symptoms and in remission. Consistent with previous reports in both rodents and humans^{31–33, 46}, the NO-mediated component of endothelium-dependent dilation was improved in SSRI-treated adults. In contrast, endothelium-independent dilation was impaired to a similar magnitude in SSRI-treated adults and MDD adults in a current depressive episode. Although these data are inherently limited by their cross-sectional nature, as well as by the potential for SSRI treatment to directly influence vascular function independent of improvements in symptoms⁶⁶, they represent a critical first step in designing future prospective studies to more specifically test research questions pertaining to the temporal association between depressive mood and endothelial dysfunction, as well as determining the mechanistic mediators of this relation.

Lastly, depression is comorbid with many cardiovascular risk factors (e.g., aging, tobacco use, obesity, physical inactivity), and it is now evident that this relation is bi-directional⁶. Over time, the added presence of any of these behavioral or physiological factors in adults with MDD would likely only exacerbate the severity of microvascular dysfunction, thereby further accelerating the atherosclerotic disease process. By study design, every attempt was made to limit these additional risk factors in the present study, in an effort to isolate the influence of MDD itself on microvascular function. However, consideration of these additional risk factors will be critical for future prospective studies designed to better understand the mechanistic link between psychological and physiological function.

Perspectives and conclusions.

These findings demonstrate that increased oxidative stress, via reductions in NO-dependent mechanisms, directly contributes to microvascular dysfunction in otherwise healthy young adults with MDD. Additionally, sex does not appear to selectively modulate NO-dependent dilatory mechanisms in depression. Collectively, these findings establish the experimental basis for conducting interventional studies targeting vascular oxidative stress as a means to improve vascular function and reduce CVD risk in adults with clinical depression. Interestingly, strategies targeting oxidative stress in the microvasculature may also have therapeutic and preventative effects on the pathogenesis of depression itself. In this regard, the ‘vascular depression’ hypothesis posits that cerebrovascular disease may predispose, precipitate, or perpetuate some depressive symptoms in the elderly⁶⁷. Based on this hypothesis, it is tempting to speculate that microvascular endothelial dysfunction, particularly in the cerebral microcirculation, could conceivably contribute to the development of depressive symptoms in otherwise healthy young adults via its effects on structural and functional neurocircuitry. This conjecture, which is supported by evidence linking inflammation—in part, via oxidative stress—to the neurobiology of depression⁵⁹, represents an exciting venue for continued research.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• In rodent models of depression, increased vascular oxidative stress, specifically superoxide production and subsequent reductions in nitric oxide (NO) bioavailability mediate endothelial dysfunction.

• Due to increased antioxidant defense mechanisms and a relative preservation of NO bioavailability, depression-induced impairments in endothelium-dependent dilation are less severe in female rodents..

• Although endothelial dysfunction is evident in adults humans with Major Depressive Disorder (MDD), the molecular mediators remain undefined.

What New Information Does This Article Contribute?

• In both men and women with MDD with mild to moderate symptomology, reductions in NO-dependent mechanisms contribute to impaired endothelium-dependent dilation; however, there are no sex differences in NO-dependent dilation in MDD.

• Acute scavenging of superoxide or inhibition of NADPH oxidase improves the NO-mediated component of endothelium-dependent dilation in MDD adults, suggesting that increased vascular oxidative stress—specifically increased superoxide—contributes to microvascular dysfunction in MDD.

Depressive disorders have been linked to the excessive and premature development of cardiovascular disease. Vascular endothelial dysfunction appears to play a pathogenic role in depression-cardiovascular disease comorbidity; however, the molecular mediators remain undefined. Based on evidence in rodent models of depression indicating a causal role for increased oxidative stress, we examined the role of superoxide in mediating microvascular endothelial dysfunction in treatment-naïve, otherwise healthy, young adults with MDD. In comparison with healthy non-depressed adults, NO-mediated endothelium-dependent dilation was attenuated in adults with MDD. Endothelium-independent dilation was also blunted.. Nevertheless, there were no differences in microvascular function between men and women with MDD. Further, in MDD, acute scavenging of superoxide or inhibition of NADPH oxidase improved NO-dependent dilation. Taken together with data demonstrating increased abundance/expression and activity of enzymes associated with an increased oxidative load, these findings support the notion that increased vascular oxidative stress contributes to microvascular dysfunction in MDD. Thus, targeting oxidative stress—specifically superoxide—may be a viable therapeutic strategy for improving vascular function and reducing cardiovascular risk in adults with depression.

Nonstandard Abbreviations and Acronyms:

Ach

acetylcholine

CVD

cardiovascular disease

eNOS

endothelial nitric oxide synthase

FMD

flow-mediated dilation

HC

healthy control

L-NAME

N^G-nitro-L-arginine methyl ester

MDD

Major Depressive Disorder

NO

nitric oxide

NOS

nitric oxide synthase

pVASP

phosphorylated vasodilator-stimulated phospho

ROS+RNS

reactive oxygen species + reactive nitrogen species

SNP

sodium nitroprusside

SOD

superoxide dismutase

SSRI

selective serotonin reuptake inhibitor

VASP

vasodilator-stimulated phosphoprotein

VSMC

vascular smooth muscle cell