Signals From Inflamed Perivascular Adipose Tissue Contribute to Small-Vessel Dysfunction in Women With Human Immunodeficiency Virus

Abstract

Background

People with the human immunodeficiency virus (PWH) have microvascular disease. Because perivascular adipose tissue (PVAT) regulates microvascular function and adipose tissue is inflamed in PWH, we tested the hypothesis that PWH have inflamed PVAT that impairs the function of their small vessels.

Methods

Subcutaneous small arteries were dissected with or without PVAT from a gluteal skin biopsy from 11 women with treated HIV (WWH) aged < 50 years and 10 matched women without HIV, and studied on isometric myographs. Nitric oxide (NO) and reactive oxygen species (ROS) were measured by fluorescence microscopy. Adipokines and markers of inflammation and ROS were assayed in PVAT.

Results

PVAT surrounding the small arteries in control women significantly (P < .05) enhanced acetylcholine-induced endothelium-dependent relaxation and NO, and reduced contractions to thromboxane and endothelin-1. However, these effects of PVAT were reduced significantly (P < .05) in WWH whose PVAT released less adiponectin but more markers of ROS and inflammation. Moderation of contractions by PVAT were correlated positively with adipose adiponectin.

Conclusions

PVAT from WWH has oxidative stress, inflammation, and reduced release of adiponectin, which may contribute to enhanced contractions and therefore could promote small-artery dysfunction.

People with the human immunodeficiency virus (HIV; PWH) have an increased risk of cardiovascular disease (CVD) [1] that is associated with impaired function of both large [2] and small arteries [3]. We have reported that small (200–250 μm) subcutaneous arteries from women with HIV (WWH), studied ex vivo on a myograph display, have marked reductions in acetylcholine (ACh)-induced endothelium-derived relaxation factor (EDRF) and nitric oxide (NO) but enhanced ACh-induced endothelium-derived contracting factor (EDCF) and reactive oxygen species (ROS), and enhanced contractions to thromboxane (U-46619) and endothelin 1 (ET-1) [4]. However, these studies were of arteries denuded of encircling perivascular adipose tissue (PVAT) that could provide a further layer of vascular regulation. This study investigates whether PVAT signaling contributes to small-vessel dysfunction in PWH in addition to the established intrinsic vascular defects.

Most blood vessels in human subjects are surrounded by PVAT [5, 6], whose signals normally enhance vasorelaxation and inhibit contraction by NO- and adiponectin-mediated pathways [5]. PVAT from obese subjects is inflamed and fails to enhance ACh-induced relaxation [7]. However, presently, it is not known if abnormal PVAT signaling contributes to microvascular dysfunction in other clinical situations of small-vessel disease such as HIV.

CVD in PWH entails increased ROS [4], inflammation [4], and endothelial dysfunction [8] that could impair PVAT function [7]. Protease inhibitors that are used to treat PWH can trigger oxidative stress and inflammation in human adipocytes and reduce adiponectin [9]. A reduced circulating level of adiponectin was associated with subclinical CVD in PWH [10]. However, a subsequent large study reported increased circulating levels of adiponectin in middle-aged WWH [11] and a U-shaped association with CVD or death has been reported in older adults [12]. Adiponectin is the most abundant adipokine in PVAT. Presently, there are no studies of the effects of HIV or its treatment on adiponectin in PVAT [13]. Although there is a complex interrelationship between the risk for CVD and HIV treatment, especially with protease inhibitors [14], studies have concluded that CVD in PWH cannot be ascribed to the effects of concurrent CVD risk factors [15] or to treatment alone [1].

Adiponectin from PVAT can exert anticontractile effects via generation of NO [16, 17] and additional antioxidative effects via recoupling of nitric oxide synthase (NOS) III [18]. Although the mass of PVAT correlates with coronary artery disease in PWH [19], the role of abnormal PVAT signaling in the small artery dysfunction of PWH has not been studied. We compared a group of young women with HIV who were free of hypertension (except that controlled by a single agent), dyslipidemia, or diabetes with a matched group of normal volunteers. Small artery function was assessed ex vivo from subcutaneous small arteries dissected from a gluteal skin biopsy and studied on an isometric myograph [4]. The role of PVAT was assessed by comparing responses of vessels surrounded by PVAT compared with denuded vessels. Endothelial function was assessed in precontracted arteries from ACh-induced EDRF, endothelium-derived hyperpolarizing factor (EDHF), and EDCF [4, 20]. Contractions to phenylephrine, U-46619, and ET-1 were assessed from dose-response curves of vessel under spontaneous tone [4, 20]. Potential mediators were assessed ex vivo in dissected PVAT.

PVAT from PWH was found to have a diminished role in enhancing ACh-induced NO generation and diminishing U-46619 and ET-1–induced contractions, which was correlated with decreased PVAT adiponectin.

METHODS

Study Population

The test group (n = 11) were self-identified premenopausal African-American women with a confirmed diagnosis of HIV for > 5 years who were enrolled in the Metropolitan Washington Women's Interagency HIV Study (WIHS), now combined with the Multicenter AIDS Cohort Study (MACS) into the MACS/WIHS Combined Cohort Study (MWCCS). Participants were selected if they had no traditional CVD risk factors except for increased body mass index (BMI), current smoking status, and mild hypertension controlled by a single drug. All had well-controlled HIV and most were currently receiving antiretroviral therapy (ART). We had planned originally to limit the study to truly virally suppressed patients with a viral load < 200 copies · mL⁻¹ but recruitment problems led us to expand the entry criteria to those with a viral load < 500 copies · mL⁻¹ within the past 3 months. The women without HIV group (n = 10) were selected from matched healthy African-American premenopausal women participating in the WIHS.

Exclusion criteria for both groups included prior stroke, myocardial infarction, kidney or liver or autoimmune disease; dementia; hypertension not controlled by a single drug; diabetes mellitus or endocrine disease; anemia; dyslipidemia; heavy alcohol use (>7 drinks per week) or current substance abuse; postmenopausal state (biologic or surgical) or receiving female sex hormones; liver function tests or serum creatinine outside the laboratory normal range; an abnormal urinalysis (except for a trace of protein); and inability to comprehend the informed consent.

Their seated blood pressure was measured after 3 minutes of rest with a mean of 3 readings. No subject had a blood pressure ≥ 140/90 mmHg or a serum creatinine ≥ 1.3 mg·dl⁻¹ or an abnormal urinalysis (except for a trace of protein) at the time of study. All were requested to refrain from smoking for 24 hours before the study and those with hypertension to refrain from taking their antihypertensives for 3 days. Each participant provided written consent as approved by the Georgetown University Institutional Review Board.

Preparation of Subcutaneous Vessels

A block of skin with subcutaneous gluteal fat (approximately 3 × 0.6 × 2 cm³) was obtained under local anesthesia and placed in physiological saline solution (PSS) at 4°C [21]. Small arteries (200 to 250 μm in diameter) were dissected and mounted on 3 isometric, 4-chamber Mulvany-Halpern small-vessel myographs (Danish MyoTech) [20] either with PVAT or stripped of PVAT [22]. The PVAT surrounding the vessel was paired down to allow the vessel to fit snugly in the myograph chamber.

Preparation of Adipose Tissue

A sample of subcutaneous fat was homogenized on ice in phosphate-buffered saline (PBS) containing antiproteases (7 mL PBS plus 1 Mini Complete Tablet; Roche), centrifuged at 14 000g at 4°C for 20 minutes, the lipid cake removed, and the adipose supernatant stored at −80°C for subsequent analysis of adipokines, cytokines, and markers of ROS and inflammation. The total protein concentrations of the tissue extracts were measured using the Pierce BCA Protein Quantitation kit.

Small Artery Protocols

The small arteries were warmed to 37°C, equilibrated for 30 minutes, and the internal circumference set to a wall tension of 0.2 mN·mm⁻² [23]. The myograph chambers were bubbled with 5% CO₂ and 21% O₂ and maintained at pH of 7.4 [21, 23]. Vessels were contracted 3 times with norepinephrine (10⁻⁵ mol·l⁻¹), followed by 1 exposure to a high-potassium solution (123 mmol·l⁻¹; KPSS) and finally a repeat exposure to KPSS containing norepinephrine (NAK). NAK provided the reference contraction [23]. Contractions were maintained for 3 minutes before rinsing with PSS. All drugs were made up in PSS and all responses compared to PSS vehicle controls. Complete studies of vessels with or without PVAT were undertaken in all WWH and women without HIV.

Ach-Induced, Endothelium-Dependent Relaxation, EDRF, EDHF, and EDCF Responses, and Sodium Nitroprusside-Induced Endothelium Independent Relaxation Responses

Vessels were preconstricted with 10⁻⁵ mol·l⁻¹ norepinephrine and relaxed with ACh (10⁻⁸ to 10⁻⁴ mol·l⁻¹ for endothelium-dependent relaxation [EDR]) or sodium nitroprusside (10⁻⁹ to 10⁻³ mol·l⁻¹ for endothelium-independent relaxation [EIR]) as described previously [4]. The EDRF response was the relaxation to ACh in PSS (vehicle) minus the same response in 10⁻⁵ mol·l⁻¹ of L-N^G-nitroarginine methyl ester (L-NAME) to inhibit NOS. The EDHF response was the relaxation to ACh in L-NAME–pretreated vessels minus the response after blockade of calcium-activated potassium channels with 10⁻⁶ mol·l⁻¹ apamin plus 10⁻⁵ mol·l⁻¹ charybdotoxin [23]. EDCF responses were the contractions to ACh of vessels under spontaneous tone after all endothelial relaxation pathways were blocked by L-NAME, apamin, and charybdotoxin.

Contractions to Phenylephrine, U-46619, and ET-1

Vessels were contracted with graded concentrations of phenylephrine (10⁻⁸ to 10⁻⁵ mol·l⁻¹), U-46619 (10⁻¹² to 10⁻⁶ mol·l⁻¹; thromboxane prostanoid receptor agonist), or ET-1 (10⁻¹⁰ to 10⁻⁷ mol·l⁻¹).

Time Control Studies

Norepinephrine contractions were similar at the beginning (85% [SEM ± 4%]) and end (78% ± 4%) of the experiments. In pilot studies, ACh relaxations were unchanged during incubation of vessels in PSS for 30, 60, and 120 minutes and remained stable over 10 hours (before, 75% ± 3% vs after 10 hours, 70% ± 5%; not significant).

Small-Artery NO and ROS Generation

NO and ROS activities of small arteries were measured as previously described [24, 25]. For studies of NO generation with ACh (10⁻⁴ mol·l), vessels were preloaded with 5 × 10⁻⁶ mol·l⁻¹ of 4-amino-5-methoxyamino-2′,7′-difluorofluorescein diacetate (DAF-FM DA; Invitrogen) and 10⁻³ mol·l⁻¹ of L-arginine. Incubation with L-NAME inhibited 93% of DAF-FM fluorescence. For studies of ROS with EDCF or ET-1 (10⁻⁷ mol·l⁻¹), vessels were preloaded with 10⁻⁵ mol·l⁻¹ 4-(9-acridinecarbonylamino)-2,2,6,6-tetramethylpiperidin-1-oxyl free radical (TEMPO-9-AC; Invitrogen) [25, 26]. Excitation was set at 360 nm and emission was isolated at 460 nm. Incubation with polyethylene glycol superoxide dismutase (125 units·mL⁻¹) prevented 92% of the EDCF fluorescence [26]. For studies of mitochondrial ROS generation, vessels were incubated with mitoSOX red (Invitrogen). This is a cell-permeant derivative of hydroethidine that targets mitochondria rapidly and selectively to emit red fluorescence (excitation/emission at λ = 510 nm/580 nm) [27].

Measurements of 8-Isoprostane F2α and Malondialdehyde

8-isoprostane F_2α in adipose supernatant was quantified by enzyme-linked immunosorbent assay (ELISA; Enzo Life Science) after purification and extraction with [³H]-Prostaglandin E₂ as an extraction marker, as validated previously. Malondialdehyde was measured by an assay kit (Cayman Chemical Company). Values were normalized per mg protein.

Measurements of Cytokines and Adipokines in Plasma and Adipose Supernatants

Measurements in plasma samples and adipose supernatants were made by Milliplex human adipokine magnetic bead panel (HADKMAG-61K and HADK2MAG-61K; Millipore Sigma) using a bead-based multiplex assay with Luminex technology at the Immunology Core, Vitalant Research Institute, San Francisco, CA, following the manufacturer's protocols.

Statistical Analysis

Data are presented as mean ± SEM. Cumulative dose-response experiments were analyzed by nonlinear regression (curve fit) and differences assessed by 2-way, repeated-measures analysis of variance (ANOVA) followed, if appropriate, with Bonferroni post hoc t tests for multiple comparisons. Associations of PVAT-induced changes in contractions with adipose levels of adiponectin and leptin were assessed by linear regression. A probability value < .05 was considered statistically significant.

RESULTS

Subject Characteristics

The 2 groups were well matched except that the estimated glomerular filtration rate (eGFR) was significantly higher in WWH than in women without HIV (Table 1). The CD4⁺ T lymphocyte count of WWH averaged 510 ± 72 cell·mL⁻³. All had reduced ART for an average of 10 years but 2 were not currently receiving therapy. They had a viral load < 500 copies· mL⁻¹; 6 had a load of < 20 copies · mL⁻¹. As expected, WWH had significantly lower CD4 counts and CD4/CD8 (P < .001) but insignificant CD8 differences. One HIV-seronegative participant and 2 WWH had hypertension that was fully controlled with a single medication, which was withheld 3 days before study. No subject had a blood pressure > 140/90 mmHg on the day of study or more than trace proteinuria and none had an abnormal serum creatinine concentration. The subjects in both groups were free of other traditional cardiovascular risk factor, except for increased BMI and current smoking status that were similar in the 2 groups. When applying multivariate regression analysis to age and HbA1c, and separately to LDL and triglycerides, as covariates in HIV status no significance was found (P = .07 and .92, respectively).

Small Arteries From WWH Have Impaired Endothelial and PVAT Functions

The PVAT-denuded preconstricted vessels from WWH had impaired ACh-induced EDR, EDRF, and EDHF and enhanced ACh-induced EDCF responses but maintained sodium nitroprusside-induced EIR (Figure 1). These results are similar to our prior report. The presence of PVAT enhanced EDR and EDRF (but not EDHF, EDCF, or EIR) significantly in small arteries from women without HIV but not in those from WWH.

Small Arteries From WWH Have Enhanced Contractions to U-46619 and ET-1 and Diminished Effect of PVAT to Moderate These Contractions

The PVAT-denuded vessels from WWH had enhanced contractions to U-46619 and ET-1 (but not to phenylephrine) and ACh (EDCF) as in our prior studies (Figure 2). The presence of PVAT significantly moderated contractions to U-46619 and ET-1 of small arteries from women without HIV but this moderating effect of PVAT was reduced in vessels from WWH.

Microvessels From WWH Have a Lessened ACh-Induced Generation of NO But Enhanced ACh- and ET-1–Induced Generation of ROS and Enhanced ET-1–Induced Generation of Mitochondrial ROS

Vessels from WWH has reduced NO generation with ACh but enhanced ROS generation with ACh or ET-1 (Figure 3A and 3B). The effect of PVAT to enhance NO was greater in vessels from women without HIV than in WWH whereas the presence of PVAT did not influence ROS generation significantly in either group. Changes in mitochondrial ROS paralleled those in cellular ROS (data not shown).

Effects of PVAT on Small Artery Function

The role of PVAT in vessels from WWH was assessed from the percentage change in vessel function induced by PVAT. PVAT from WWH had a significantly diminished effect in both augmenting NO generation with ACh and moderating contractions with U-46619 or ET-1 but other microvascular functions were not significantly affected (Figure 4).

Mediators in PVAT and Plasma

Compared to controls, PVAT from WWH had a reduced adiponectin, but an increased leptin and resistin, and increased markers of inflammation (tumor necrosis factor-α [TNF-α], plasminogen activator inhibitor-1 [PAI-1], and nerve growth factor [NGF]), and ROS (8-isoprostane F₂α and malondialdehyde) (Table 2). Plasma from WWH showed these same changes, except that plasma resistin, malondialdehyde, and TNF-α were not different between groups. In addition, plasma from WWH had decreased nitric oxide metabolites.

Correlation of PVAT-Induced Vessel Function With PVAT Adipokines

PVAT-induced reductions in microvascular contractions correlated positively with adipose levels of adiponectin among all participants. Reductions in contractions to U-46619 and ET-1 by PVAT correlated positively and significantly with adipose levels of adiponectin (Figure 5A and 5B) but not with adipose levels of leptin (Figure 5C and 5D) or resistin (data not shown).

DISCUSSION

We confirm our previous report that denuded subcutaneous small arteries from young women with HIV have impaired intrinsic NO generation and endothelial relaxation functions and enhanced ROS generation and contractions to U-46619 and ET-1 (but not to phenylephrine) [4]. The present study is the first to demonstrate additional microvascular defects that are related specifically to PVAT. Thus, the PVAT-induced enhanced NO generation and reduced contractions to U-46619 and ET-1 in small arteries of women without HIV were lessened in vessels from WWH. PVAT from WWH released more leptin and markers of inflammation and ROS, but less adiponectin. The decreased release of adiponectin from PVAT of WWH could have contributed to the enhanced contractions because the effects of PVAT to reduce these contractions correlated positively with adipose levels of adiponectin [28–32]. We conclude that PVAT from young women with HIV has inflammation, oxidative stress, and reduced release of adipokines, which may contribute to a reduced NO generation and enhanced contractility that may result in increased risk for CVD.

PVAT generally serves to regulate vascular function and protect blood vessels from CVD. Factors released from PVAT can act directly on vascular smooth muscle cells (VSMCs). PVAT itself can generate NO [5] or induce vascular endothelial cells to generate NO that activates guanosine monophosphate kinase to relax VSMC [33]. Release of NO from small vessels surrounded by PVAT derived from normal subjects was greater than from corresponding PVAT-denuded vessels. However, this does not differentiate between NO generated in PVAT and PVAT-induced augmentation of NO generated in the endothelium of the small arteries. PVAT can release adiponectin that induces small-vessel relaxation. PVAT also can enhance EDHF in some experimental studies [34]. However, we did not detect a significant effect of PVAT on EDHF or EDCF responses in this study.

The potential role of adipokines was studied in samples of PVAT ex vivo. The reduction in PVAT adiponectin in WWH might contribute to impaired NO generation because adiponectin can recouple and phosphorylate NOS-3 and increase the bioavailability of the NOS cofactor, tetrahydrobiopterin [18]. Thus, the novel finding of a reduction in PVAT-derived adiponectin in WWH in this study could have contributed to the reduced release of NO by small arteries. Circulating adiponectin could not be related to the function of PVAT that was studied ex vivo. It would be interesting in a future study to relate circulating and PVAT adiponectin to PVAT function in samples gathered at the same time. In turn, an increased production of ROS in inflamed small arteries from WWH could have induced inflammation and increased interleukin 1b (IL-1b) [35], TNF-α [36], PAI-1 [37], and NGF [38]. However, ROS generation was not dependent on PVAT in this study and PVAT did not contribute to EDCF responses that depend on ROS [22].

We acknowledge some limitations of our study. First, the number of participants is insufficient to exclude a type 2 error that prevents us from concluding that a factor was or was not different between the groups. Thus, this study should be considered as a hypothesis-generating pilot study. In this somewhat invasive first study of PVAT function in WWH, we selected a group of subjects of one sex (female), one ethnicity (African American), and limited age span (20–50 years) in the premenopausal state with little CVD risk. This yielded a homogenous group to study but limited the available numbers. Second, because correlation cannot prove causality, subsequent intervention studies will be required to establish a causal role for PVAT adiponectin. Third, the quantity of PVAT surrounding each vessel was not controlled rigorously but was approximated between vessels by the quantity that permitted the vessel to fit snugly into its bath. Fourth, studies were confined to subdermal blood vessels. However, microvascular dysfunction is now considered a systemic disease that is worsened by vascular risk factors and represents a global pathological process [39].

In conclusion, PVAT surrounding small arteries from WWH has a diminished capacity to release NO, which could contribute to its diminished capacity to moderate contractions to U-46619 and ET-1. A decreased generation of adiponectin in PVAT from WWH could contribute to the attenuated NO-generating effect of PVAT in WWH and thereby to its attenuated anticontractile effects.

Our results suggest that PVAT from WWH has a reduced anticontractile effect that is independent of ROS because the increased arteriolar ROS in WWH was unaffected by the presence of PVAT. Rather, PVAT from WWH had a diminished effect in stimulating NO generation with ACh. Angiotensin II [40], thromboxane [41], and ET-1 [42] release vascular NO that offsets their vasoconstrictor actions on VSMCs. Thus, a reduced release of NO, rather than enhanced ROS, may be of predominant importance for the impaired anticontractile effects of PVAT from WWH. In turn, a reduced release of NO in WWH could be a consequence of the reduced release of adiponectin from PVAT [16, 18, 22]. Any ensuing enhancement of small-vessel contractility could contribute to vascular dysfunction and perhaps to hypertension that is common in WWH [1]. Although PVAT from WWH had increased leptin release, the consequence is unclear because leptin can enhance NO generation in some studies [43], but can mediate endothelial dysfunction in others [44, 45] by mechanisms that are presently poorly characterized [46]. Our results favor a predominant role for reduced PVAT adiponectin rather than enhanced PVAT leptin in the attenuated anticontractile effects of PVAT from WWH because these correlated quite strongly with reduced adiponectin but not significantly with leptin (Figure 5). Further studies will be required to explore these possibilities.

Notes

Acknowledgments. We thank staffs of the Women's Interagency HIV Study (WIHS) and Division of Infection Disease, Department of Medicine, Georgetown University Hospital for recruitment of follow-up of participants, and the patients for collaboration; Jennifer Verbasey, MD for undertaking the gluteal skin biopsies; and the Georgetown University Clinical Research Unit and the Clinical and Translational Science Program for clinical facilities, statistical advice, and assistance with gluteal biopsy.

Financial support. This work was supported by the National Institutes of Health (NIH) (grant number R01 HL13451101 to D. W. and C. S. W.); the Smith-Kogod Family Foundation (to C. S. W.); the Gildenhorn-Spiesman Family Foundation; the Georgetown University Hypertension Center; the Schreiner Chair of Nephrology (to C. S. W.); the Walters Family Chair of Cardiovascular Research (to C. S. W.); the Marriott Family Research Award (to D. W.); National Heart, Lung, and Blood Institute (grant number R01 HL124511); National Institute of Allergy and Infectious Diseases American Recovery and Reinvestment Act grant (grant numbers U01 A1034994 to WIHS and U01 HL146205 to the Multicenter AIDS Cohort Study/WIHS Combined Cohort Study/Metropolitan Washington Clinical Research Center); District of Columbia Center for AIDS Research, an NIH funded program (supplemental award grant number P30AI117970); National Center for Advancing Translational Science Federal fund (grant number UL1 TR000101); NIH Clinical and Translational Science Awards Program, part of the “Roadmap Initiative, Re-Engineering the Clinical Research Enterprise”; and Department of Medicine Georgetown University Translational Research Pilot Award (to D. W.).

References

Author notes