Increased peripheral vascular disease risk progressively constrains perfusion adaptability in the skeletal muscle microcirculation

Linking peripheral vascular disease (PVD) risk to integrated microvascular dysfunction and health outcomes has been elusive. We used eight models of increasing risk and a multiscale approach to interrogate novel indexes of microvascular function, perfusion/oxygen handling, and outcomes. We demonstrate how elevated PVD risk leads to progressive microvascular “dampening,” with clear implications for outcomes.

Abstract

To determine the impact of progressive elevations in peripheral vascular disease (PVD) risk on microvascular function, we utilized eight rat models spanning “healthy” to “high PVD risk” and used a multiscale approach to interrogate microvascular function and outcomes: healthy: Sprague-Dawley rats (SDR) and lean Zucker rats (LZR); mild risk: SDR on high-salt diet (HSD) and SDR on high-fructose diet (HFD); moderate risk: reduced renal mass-hypertensive rats (RRM) and spontaneously hypertensive rats (SHR); high risk: obese Zucker rats (OZR) and Dahl salt-sensitive rats (DSS). Vascular reactivity and biochemical analyses demonstrated that even mild elevations in PVD risk severely attenuated nitric oxide (NO) bioavailability and caused progressive shifts in arachidonic acid metabolism, increasing thromboxane A₂ levels. With the introduction of hypertension, arteriolar myogenic activation and adrenergic constriction were increased. However, while functional hyperemia and fatigue resistance of in situ skeletal muscle were not impacted with mild or moderate PVD risk, blood oxygen handling suggested an increasingly heterogeneous perfusion within resting and contracting skeletal muscle. Analysis of in situ networks demonstrated an increasingly stable and heterogeneous distribution of perfusion at arteriolar bifurcations with elevated PVD risk, a phenomenon that was manifested first in the distal microcirculation and evolved proximally with increasing risk. The increased perfusion distribution heterogeneity and loss of flexibility throughout the microvascular network, the result of the combined effects on NO bioavailability, arachidonic acid metabolism, myogenic activation, and adrenergic constriction, may represent the most accurate predictor of the skeletal muscle microvasculopathy and poor health outcomes associated with chronic elevations in PVD risk.

NEW & NOTEWORTHY

Linking peripheral vascular disease (PVD) risk to integrated microvascular dysfunction and health outcomes has been elusive. We used eight models of increasing risk and a multiscale approach to interrogate novel indexes of microvascular function, perfusion/oxygen handling, and outcomes. We demonstrate how elevated PVD risk leads to progressive microvascular “dampening,” with clear implications for outcomes.

the predominant concern regarding the presence of peripheral vascular disease (PVD) risk factors of increasing severity is that these can lead to the development of pathological characteristics of PVD including myocardial infarction, heart failure, stoke, and/or limb ischemia. These potential outcomes of PVD result in elevations in mortality risk as well as significant reductions in quality of life through a variety of direct and indirect causes (3, 45, 50) that are predominantly perceived as macrovascular events. These global processes can develop across tissues and organs, leading to the clinical manifestations of ischemic episodes/stroke in the brain, angina/infarction in the heart, and claudication/obstruction in the skeletal muscle circulation, among others. However, these macrovascular events do not occur in isolation from significant microvascular derangement.

With specific attention to the skeletal muscle circulation and its role in the development of PVD, there has been an enormous accumulation of knowledge regarding associations between elevated PVD risk, specific structural or functional markers of vascular dysfunction, putative mechanistic contributors and indexes of perfusion-based deficits, and impaired tissue/organ function (11, 25, 26, 38, 46). While this previous work has shed considerable light on what is a very complicated issue, it has proven very difficult to assign and understand the multiscale nature of PVD risk and negative vascular outcomes both within and across models of pathology. In a nutshell, an accurate understanding of how “endothelial dysfunction” or “adrenergic constriction” (as examples) contributes not only to dysregulation of vascular resistance and perfusion but also to the accelerated development of skeletal muscle fatigue with increased metabolic demand has proven elusive. While there is a compelling argument that increasing PVD risk and disease severity across models may not be directly comparable owing to the diversity of specific risk factors and their signaling pathways, it is more biologically plausible that there are critical underlying factors and common pathways that largely define “peripheral vascular disease” and poor health outcomes under this condition.

The obese Zucker rat (OZR) is a translationally relevant model of the metabolic syndrome, with its basis in chronic hyperphagia owing to leptin resistance, and the ensuing development of obesity, impaired glycemic control, dyslipidemia, and moderate hypertension (1, 18, 23, 53). The OZR is a model of significant PVD risk (in the absence of significant atherosclerosis), and previous results have identified a series of vasculopathies and contributing mechanisms that can impair skeletal muscle perfusion and performance (20, 24, 61, 62). Going forward, recent studies have shed new light on microvascular dysfunction itself that may represent a more accurate predictor of poor outcomes in OZR than has been available previously. Specifically, previous studies have identified an attractor that integrates vascular hemodynamics and reactivity at successive arteriolar bifurcations to create a condition in which terminal arteriolar perfusion becomes both increasingly heterogeneous and less responsive to changes in the local environment (5, 21).

An attractor may be best described as a graphic representation of the set of behaviors toward which a dynamic system evolves over time (40, 57). This has the potential of being very informative in understanding the limits of a system (i.e., the likelihood of a dynamic system assuming a particular state at a moment in time) and the effects on system behavior (either beneficial or detrimental) subsequent to an imposed intervention or challenge. Attractors are generally presented as a three-dimensional iterated map, where the current state of the parameter/system under study is plotted and then its change in position replotted at the next time interval. With sufficient iteration (for our purposes, to the termination of the collected data set), the shape and location of the attractor on the coordinate system become evident, facilitating the ability to make comparisons between experimental conditions. These comparisons would be made through a change in the shape and/or position of the attractor (40, 57).

This outcome in OZR appears to reflect a spatial and temporal integration of alterations to nitric oxide (NO) bioavailability/oxidant stress, altered arachidonic acid metabolism, increased myogenic activation, and a shift in adrenergic-based constrictor reactivity, creating an environment characterized by compromised matching of perfusion to metabolic demand and premature muscle fatigue (22, 32, 41). However, to date this has been identified in only one model (OZR), and it is unknown whether this shift in attractors describing microvascular behavior is present across multiple models of PVD risk with different constitutions. For example, are these effects seen in isolation in high-PVD risk, metabolically deranged models, or are these effects a graded response in correlation with increasing PVD risk factor severity?

The purpose of the present study was to begin to determine the applicability of this approach across models of elevated PVD risk of varied severity. We have employed eight models ranging from “healthy/minimal PVD risk” to “high PVD risk” and have determined relationships between basic indexes of vascular function and functional hyperemic responses and muscle performance. The present study was extended to determine the attractor for microvascular perfusion distribution in both the proximal and distal arteriolar bifurcations within the microvascular networks within skeletal muscle, with a prediction of precapillary arteriolar network perfusion distribution and flexibility with increasing PVD risk. The general hypothesis tested by the present study was that with increasing severity of PVD risk the attractor describing microvascular perfusion distribution at successive arteriolar bifurcations is increasingly displaced from a position of optimal responsiveness and becomes increasingly constrained to unequal perfusion distribution with a progressive loss of system flexibility and adaptability.

MATERIALS AND METHODS

Animals

All experiments were performed in male rats ∼17 wk of age, and, unless otherwise stated (below), all animals were fed standard chow and tap water ad libitum throughout. All rats were housed in the animal care facilities at either the West Virginia University Health Sciences Center or the Medical College of Wisconsin, and all protocols received prior Institutional Animal Care and Use Committee (IACUC) approval. Prior to final usage, all rats in the present study were subjected to an overnight fast of ∼8-h duration. In all groups of animals, n = 5–7 for experimental protocol 1 and experimental protocol 2 and n = 6–10 for experimental protocol 3. The animal groups used for the present study were as follows: 1) Sprague-Dawley rats (SDR; healthy control group 1); 2) lean Zucker rats (LZR; healthy control group 2); 3) SDR on high-salt diet [HSD; 4.0% NaCl in chow for 4 wk prior to use (57); mild PVD risk group 1]; 4) SDR on high-fructose diet [HFD; 66% fructose in diet for 8 wk prior to use (12, 49); mild PVD risk group 2]; 5) reduced renal mass-hypertensive rats [RRM; SDR after receiving renal mass reduction surgery and 4.0% NaCl in chow for 4 wk prior to use (31, 39); moderate PVD risk group 1]; 6) spontaneously hypertensive rats (SHR; moderate PVD risk group 2); 7) OZR (high PVD risk group 1); 8) Dahl salt-sensitive rats [DSS; 4.0% NaCl in chow for 4 wk prior to use (17, 36); high PVD risk group 2].

Calculation of PVD Risk Score

An estimated PVD risk score for each animal group was calculated based on the severity of individual constituents listed below. Each parameter was scored ranging from 0 to 4, inclusive, with “normal” defined as the values determined for SDR and LZR. These were determined as follows: 1) Body mass (obesity) was binned by +25% ranges above normal. 2) Blood glucose concentration (insulin resistance) was binned by +25% in fasting glucose over normal. 3) Mean arterial pressure (hypertension) was binned by +20% ranges above normal. 4) Plasma cholesterol concentration (dyslipidemia) was binned by +25% ranges above normal. 5) Plasma TNF-α concentration (chronic inflammation) was binned by +250% ranges above normal. 6) Plasma nitrotyrosine concentration (oxidant stress) was binned by +100% ranges above normal. The total of the six individual scores was determined as an estimator of the aggregate PVD risk within an animal group. While this provides some estimation of PVD risk severity across the different groups, it is obviously not exhaustive or complete, as it contains no parameter weighting.

Experimental Protocol 1 (Vascular Reactivity)

At the time of final usage, each rat was anesthetized with injections of pentobarbital sodium (50 mg/kg ip), and all rats received tracheal intubation to facilitate maintenance of a patent airway. In all rats, a carotid artery and an external jugular vein were cannulated for determination of arterial pressure and for intravenous infusion of additional substances as necessary (e.g., anesthetic, heparin). In addition, an aliquot of mixed venous blood was drawn from the jugular vein cannula for profiling of metabolic, endocrine, and inflammatory biomarkers (see below).

Preparation of isolated skeletal muscle resistance arterioles.

In all rats, the intramuscular continuation of the gracilis artery from each leg was removed and cannulated as described previously (39). Arterioles were doubly cannulated and placed in a heated chamber (37°C) that allowed the vessel lumen and exterior to be perfused and superfused, respectively, with physiological salt solution (PSS; equilibrated with 21% O₂, 5% CO₂, 74% N₂) from separate reservoirs. Vessel diameter was measured with television microscopy and an on-screen video micrometer. Arterioles were extended to their in situ length and were equilibrated at 80% of the animal's mean arterial pressure (19).

Within each arteriole, vessel reactivity was evaluated in response to application of acetylcholine (10⁻⁹–10⁻⁶ M) and phenylephrine (10⁻¹⁰–10⁻⁷ M) and alterations in intraluminal pressure (myogenic activation; from 40 mmHg to 160 mmHg in randomized 20-mmHg increments). Subsequently, vessels were treated with TEMPOL (10⁻⁴ M) and/or nitro-l-arginine methyl ester (l-NAME; 10⁻⁴ M) to assess the contribution of vascular oxidant stress and endothelium-dependent NO production to these mechanical responses.

Preparation of in situ cremaster muscle.

After removal of the gracilis arterioles, an in situ cremaster muscle was prepared for intravital microscopy (21) and was continuously superfused with PSS equilibrated with a 5% CO₂-95% N₂ gas mixture, maintained at 35°C as it flowed over the muscle. Volume flow rate was ∼3.0 ml/min. The ionic composition of the PSS was as follows (mM): 119.0 NaCl, 4.7 KCl, 1.6 CaCl₂, 1.18 NaH₂PO₄, 1.17 MgSO₄, 24.0 NaHCO₃, and 0.03 disodium EDTA. After a 30-min period of equilibration following the surgical preparation, distal arterioles of ∼30-μm diameter were selected for investigation based on the following criteria: 1) distance from any site of incision, 2) presence of significant vascular tone (assessed by brisk dilator response to challenge with 10⁻³ M adenosine), and 3) clearly discernible walls.

Distal arteriolar (3A-4A) dilator responses (by on-screen videomicroscopy) were determined in response to challenge with acetylcholine (10⁻⁹–10⁻⁶ M) or arachidonic acid (10⁻⁹–10⁻⁶ M) under untreated (control) conditions and after pretreatment of the cremaster muscle with TEMPOL (10⁻⁴ M; both), l-NAME (10⁻⁴ M; acetylcholine only), indomethacin (10⁻⁵ M; arachidonic acid only), or SQ-29548 (10⁻⁵ M; arachidonic acid only).

Measurement of vascular NO bioavailability.

After completion of the cremaster arteriolar reactivity experiments, the abdominal aorta was removed and vascular NO production was assessed with amperometric sensors (World Precision Instruments, Sarasota, FL). Briefly, aortas were isolated, sectioned longitudinally, pinned in a Silastic-coated dish, and superfused with warmed (37°C) PSS equilibrated with 95% O₂-5% CO₂. An NO sensor (ISO-NOPF 100) was placed in close apposition to the endothelial surface, and a baseline level of current was obtained. Subsequently, increasing concentrations of methacholine (10⁻¹⁰–10⁻⁶ M) were added to the bath and the changes in current were determined (8). To verify that responses represented NO release, these procedures were repeated after pretreatment of the aortic strip with l-NAME (10⁻⁴ M).

Determination of vascular metabolites of arachidonic acid.

Vascular production of 6-keto-prostaglandin F_1α (6-keto-PGF_1α, the stable breakdown product of PGI₂; Refs. 39, 44) and 11-dehydro-thromboxane B₂ [11-dehydro-TxB₂, the stable plasma breakdown product of thromboxane A₂ (TxA₂); Ref. 7] in response to challenge with 10⁻⁶ M arachidonic acid was assessed in pooled conduit arteries (femoral, saphenous, iliac) from each rat. Pooled arteries from each animal were incubated in microcentrifuge tubes in 1 ml of PSS for 30 min under control conditions (21% O₂). After this time, the superfusate was removed, stored in a new microcentrifuge tube, and frozen in liquid N₂, while a new aliquot of PSS was added to the vessels and the arachidonic acid was added for the subsequent 30 min. After the second 30-min period, this new PSS was transferred to a fresh tube, frozen in liquid N₂, and stored at −80°C. Metabolite release by the vessels was determined with commercially available enzyme immunoassay (EIA) kits for 6-keto-PGF_1α and 11-dehydro-TxB₂ (Cayman).

Analyses of data.

Any cannulated arteriole that did not demonstrate significant active tone at the equilibration pressure was discarded. Active tone at the equilibration pressure was calculated as (ΔD/D_max) × 100, where ΔD is the diameter increase from rest in response to Ca²⁺-free PSS and D_max is the maximum diameter measured at the equilibration pressure in Ca²⁺-free PSS. Data describing ex vivo arteriolar active tone in the present study are summarized in Table 1.

Table 1.

Baseline characteristics of animal groups

	SDR	LZR	HSD	HFD	RRM	SHR	OZR	DSS
Mass, g	344 ± 7	365 ± 8	348 ± 9	374 ± 7	345 ± 8	409 ± 6*	691 ± 12*	446 ± 8*
MAP, mmHg	101 ± 6	105 ± 5	105 ± 6	108 ± 5	161 ± 8*	164 ± 6*	132 ± 6*	189 ± 6*
GlucoseBlood, mg/dl	84 ± 4	88 ± 7	90 ± 5	113 ± 6*	94 ± 7	116 ± 6*	164 ± 8*	132 ± 7*
InsulinPlasma, ng/ml	0.8 ± 0.2	1.0 ± 0.3	0.8 ± 0.3	5.8 ± 0.7*	1.9 ± 0.3*	2.4 ± 0.6*	6.7 ± 0.6*	3.1 ± 0.7*
CholestPlasma, mg/dl	90 ± 5	92 ± 7	89 ± 6	95 ± 7	91 ± 5	109 ± 6*	124 ± 7*	122 ± 6*
TriGlyPlasma, mg/dl	84 ± 8	95 ± 10	94 ± 8	208 ± 14*	109 ± 10	170 ± 13*	348 ± 28*	276 ± 24*
TNF-αPlasma, pg/ml	0.5 ± 0.1	0.7 ± 0.3	1.9 ± 0.3*	2.2 ± 0.5*	3.2 ± 0.6*	4.1 ± 0.5*	7.1 ± 0.8*	7.4 ± 0.7*
N-tyrosinePlasma, ng/ml	14 ± 3	15 ± 4	22 ± 4	28 ± 5*	36 ± 5*	42 ± 6*	56 ± 6*	53 ± 7*
Vascular active tone, %	28 ± 3	30 ± 3	27 ± 4	31 ± 4	35 ± 4	36 ± 5	33 ± 4	37 ± 5
Q̇Muscle, ml·g−1·min−1	0.13 ± 0.02	0.12 ± 0.02	0.12 ± 0.01	0.13 ± 0.01	0.11 ± 0.02	0.10 ± 0.01	0.08 ± 0.02*	0.07 ± 0.02*
Peak force, g	978 ± 43	966 ± 53	981 ± 51	972 ± 49	958 ± 52	965 ± 54	956 ± 56	950 ± 44

Values are means ± SE. SDR, Sprague-Dawley rats; LZR, lean Zucker rats; HSD, SDR on high-salt diet; HFD, SDR on high-fructose diet; RRM, reduced renal mass-hypertensive SDR; SHR, spontaneously hypertensive rats; OZR, obese Zucker rats; DSS, Dahl salt-sensitive rats; MAP, mean arterial pressure; Q̇_Muscle, arterial blood flow.

^*P < 0.05 vs. LZR.

The mechanical responses of isolated and in situ arterioles following pharmacological challenge (acetylcholine, phenylephrine, or arachidonic acid) were fit with the three-parameter logistic equation:

$$y=\min +\left[\frac{\max -\min }{1+{10}^{\log {\text{ED}}_{50}-x}}\right]$$

where y represents the change in arteriolar diameter, “min” and “max” represent the lower and upper bounds, respectively, of the change in arteriolar diameter with increasing agonist concentration, x is the logarithm of the agonist concentration, and logED₅₀ represents the logarithm of the agonist concentration (x) at which the response (y) is halfway between the lower and upper bounds.

The myogenic activation for each experimental group was plotted as mean diameter at each intraluminal pressure and fitted with a linear regression (y = α₀ + βx), where y represents vessel diameter, x represents intraluminal pressure, α₀ represents an intercept term, and the slope coefficient β represents the degree of myogenic activation (δdiameter/δpressure). Increasingly negative values of β therefore represent a greater degree of myogenic activation in response to changes in intraluminal pressure. A similar analysis was used to determine the NO bioavailability in response to increasing concentrations of methacholine, where β represents the rate of change of NO released by the vessels in response to agonist challenge.

Experimental Protocol 2 (In Situ Skeletal Muscle Contraction and Perfusion)

Preparation of in situ blood-perfused skeletal muscle.

After the removal of a gracilis muscle resistance arteriole, as also described for the animals in experimental protocol 1, the left gastrocnemius muscle was isolated in situ (20). Briefly, the left leg received a medial incision from the calcaneus to the femoral triangle, and all muscles, vessels, and connective tissue overlying gastrocnemius muscle were removed, thus exposing the gastrocnemius muscle, its vascular supply, and the sciatic nerve. The nerve was isolated and utilized for initiating muscle contraction via a stimulating electrode attached to an electrical stimulator (Grass SD9). Branches from the femoral artery that did not perfuse the gastrocnemius muscle directly were ligated or cauterized, depending on size and location. A microcirculation flow probe (Transonic) was placed around the femoral artery, immediately distal to its origin from the iliac artery, in order to measure blood flow to the gastrocnemius muscle. As a final step, a 24-gauge angiocatheter was inserted into the femoral vein to allow sampling of venous blood from the contracting muscle. The entire preparation was covered in PSS-soaked gauze and plastic film to minimize evaporative water loss and was placed under a lamp to maintain temperature at 37°C. At this time, heparin (500 U/kg) was infused via the jugular vein to prevent blood coagulation.

After an equilibration period of ∼30 min, the gastrocnemius muscle was set to its optimal length for tension development (L_o) and was stimulated to perform (via the sciatic nerve) bouts of isometric twitch contractions (3 or 5 Hz, 0.4-ms duration, 5 V) for 3 min followed by 15 min of self-perfused recovery, with arterial pressure and femoral artery blood flow continuously monitored.

At the conclusion of the gastrocnemius contraction/perfusion experiments, conduit arteries were harvested and used for the assessment of NO bioavailability and arachidonic acid metabolites in the identical manner as described for experimental protocol 1 above.

Analyses of data.

During the final 30 s of the contraction bout (3 or 5 Hz), 200-μl blood samples were drawn from the carotid artery cannula and femoral vein angiocatheter. Samples were briefly stored on ice until they were processed for blood gas pressures, percent oxygen saturation, and hemoglobin concentration with a Corning Rapidlab 248 blood gas analyzer. Muscle perfusion, arterial pressure, and bulk blood flow through the femoral artery were monitored for 1 min prior to muscle contractions and throughout the contraction period with a Biopac MP150 and Acqknowledge data acquisition software at a 50-Hz sampling frequency. Muscle perfusion and performance data after 3 min of contraction were normalized to gastrocnemius mass, which was not different between LZR (2.22 ± 0.08 g) and OZR (2.14 ± 0.09 g). Oxygen content within the blood samples was determined with the following standard equation:

$$\text{C}{x}_{{\mathrm{\text{O}}}_{\mathrm{2}}}=\left(1.39\times \left[\text{Hb}\right]\times \%{\text{SatO}}_2\right)+\left(0.003\times \text{P}{x}_{{\mathrm{\text{O}}}_{\mathrm{2}}}\right)$$ (1)

where Cx_O2 and Px_O2 represent the total content (ml/dl) and partial pressure of oxygen (mmHg), respectively, of arterial or venous blood (denoted simply as x). [Hb] represents hemoglobin concentration within the blood sample (g/dl); %SatO₂ represents the percent oxygen saturation of the hemoglobin, and 1.39 and 0.003 represent constants describing the amounts of bound and dissolved oxygen in blood. Oxygen consumption across the gastrocnemius muscle was calculated with the Fick equation:

$$\stackrel{\cdot }{{\displaystyle \mathrm{\text{V}}}}{\mathrm{\text{O}}}_2=\stackrel{\cdot }{{\displaystyle \mathrm{\text{Q}}}}\times \left(\text{C}{\mathrm{a}}_{{\mathrm{\text{O}}}_{\mathrm{2}}}-\text{C}{\mathrm{v}}_{{\mathrm{\text{O}}}_{\mathrm{2}}}\right)$$ (2)

where V̇o₂ represents oxygen consumption by the gastrocnemius muscle, Q̇ represents femoral artery blood flow (ml·g⁻¹·min⁻¹), and Ca_O2 and Cv_O2 represent arterial and venous oxygen content, respectively.

Experimental Protocol 3 (In Situ Microvascular Network Perfusion)

Preparation of in situ cremaster muscle.

In each rat, an in situ cremaster muscle was prepared for intravital microscopy as described above. After a 30-min period of equilibration after surgical preparation, arterioles and bifurcations within five distinct diameter categories were selected for investigation, ∼100 μm (1A), ∼80 μm (2A), ∼60 μm (3A), ∼40 μm (4A), and ∼20 μm (5A) in diameter. This resulted in four “bifurcation categories”: 1) ∼100-μm “parent” to ∼80-μm “daughters” (1A-2A), 2) ∼80-μm “parent” to ∼60-μm “daughters” (2A-3A), 3) ∼60-μm “parent” to ∼40-μm “daughters” (3A-4A), and 4) ∼40-μm “parent” to ∼20-μm “daughters” (4A-5A). Arterioles/bifurcations were selected based on the following criteria: 1) distance from any site of incision, 2) presence of significant vascular tone (assessed by brisk dilator response to challenge with 10⁻³ M adenosine), 3) clearly discernible walls, 4) a rapid and stable level of erythrocyte perfusion, and 5) the presence of two clearly defined “daughter” branches that also met criteria 1–4. See Fig. 1 for a schematic representation (also Refs. 5 and 21).

Fig. 1.

Schematic representation of the in situ cremasteric arteriolar bifurcation used for assessing “parent” and “daughter” arteriolar mechanical and hemodynamic/perfusion responses to pharmacological challenge. Open arrows represent parent or daughter arteriolar diameter in response to a specific condition; filled arrows represent parent or daughter arteriolar erythrocyte (RBC) velocity in response to a specific challenge. These data are utilized to determine both arteriolar flow volume and perfusion heterogeneity at bifurcations (γ). See text for additional details, as well as Refs. 5 and 21.

Within the cremaster muscle of each animal, the mechanical (by on-screen videomicroscopy) and perfusion (with an optical Doppler velocimeter; Microcirculation Research Institute, Texas A&M University, College Station, TX) responses of both the “parent” and “daughter” arterioles within a bifurcation class were determined.

Investigation of arteriolar bifurcations.

Within each animal, one of each of the four classifications of arteriolar bifurcations was studied within the arteriolar network. For an individual measurement, arteriolar diameter and erythrocyte velocity were monitored for 5 min, with measurements taken every 20 s. All arterioles and bifurcations were selected on the basis of the criteria outlined above and were placed into their categories by size rather than by strict branch number within the network.

To maintain preparation and collected data quality, maximum experimental duration from preparation to termination was ∼3 h, after which time all animals were humanely euthanized by an intravenous overdose of anesthetic followed by a bilateral pneumothoracotomy, as per approved IACUC protocols.

Analyses of data.

Arteriolar perfusion in both parent and daughter vessels within the in situ cremaster muscle of LZR and OZR was calculated as

$$\stackrel{\cdot }{{\displaystyle \mathrm{\text{Q}}}}=\left(V\times {1.6}^{-1}\right)\left(\pi r^2\right)\left(0.001\right)$$

where Q̇ represents arteriolar perfusion (nl/s), V represents the measured red blood cell velocity from the optical Doppler velocimeter (mm/s; with V/1.6 representing an estimated average velocity assuming a parabolic flow profile; Ref. 15), and r represents arteriolar radius (μm; Ref. 2).

The total volume perfusion in the daughter arterioles was determined as the sum of the individual perfusion rates, and the proportion of flow within each was determined as the quotient of the individual branch divided by the total. γ is defined as the ratio of the greater of the two flows in the daughter vessel to the total flow in the parent vessel. As an example, if flow distribution was homogeneous between daughters, γ for that bifurcation would be 0.5 in both daughter arterioles, while if the proportion of flow in one daughter arteriole was 60%, γ for that bifurcation would be 0.6, with flow distribution being 0.6 in the “high-perfusion” arteriole and 0.4 in the “low-perfusion” arteriole (Fig. 1).

Monitoring γ throughout the observation period and analyzing the resulting data used three initial approaches. The first was simply to determine the mean ± SE of the individual measurements throughout the 5-min observation window, providing the mean value for γ. However, this is clearly not sufficiently informative in terms of understanding the temporal behavior for γ at a bifurcation. The second approach was to determine the cumulative change in γ from each of the 15 successive measurements:

$${\displaystyle \sum _{t=1}^{15}}\left({\gamma }_{\left(t\right)}-{\gamma }_{\left(t+1\right)}\right)$$

where γ represents the perfusion distribution at any given bifurcation and t represents the measurement time from 1 (initial) to 15 (final). While these data can provide us with evidence of broader changes in γ with longer durations, it was apparent that a more sensitive marker is needed to provide sufficient insight into temporal switching at an arteriolar bifurcation. As such, the final approach is to summate the absolute values of the differences in γ from one time point to the next to provide a superior index of total activity:

$${\displaystyle \sum _{t=1}^{15}}\text{ABS}\left\{\left({\gamma }_{\left(t\right)}-{\gamma }_{\left(t+1\right)}\right)\right\}$$

This final index (summating absolute difference in γ) provides insight into the total temporal switching at any arteriolar bifurcation within the data collection window.

Statistical Analyses

All data are presented as means ± SE. Differences in calculated parameters (e.g., bounds or slopes of stimulus-response curves, calculated blood gas values, blood flow, muscle force development) were assessed by analysis of variance (ANOVA) or regression analyses, with Student-Newman-Keuls test post hoc, as appropriate. In all cases, P < 0.05 was taken to reflect statistical significance.

RESULTS

Data describing the baseline characteristics of the animal groups used in the present study are summarized in Table 1. SDR and LZR were used as the “healthy control” animals, and, with the exception of LZR being somewhat larger than SDR, both strains demonstrate responses that are indicative of being termed healthy at ∼17 wk of age. HSD and HFD exhibited a mild elevation in PVD risk compared with SDR and LZR, with some evidence for the development of a prooxidant and proinflammatory state (for both) and significant insulin resistance (for HFD only). RRM and SHR demonstrated a greater PVD risk profile, with increasing severity of hypertension, chronic inflammation, and a more severe prooxidant condition with a mild insulin resistance and dyslipidemia. Finally, OZR and DSS exhibited a more severe state of PVD risk, as would be predicted by models of the full metabolic syndrome. While OZR were more biased toward obesity and impaired glycemic control, the severe hypertension in DSS, combined with the state of inflammation and oxidant stress, resulted in an estimated PVD risk score that approximated that of OZR. A graphical representation of estimated PVD risk score is presented in Fig. 2. The active tone of ex vivo gracilis arterioles was not significantly different between strains, although there was a trend toward elevated tone in models manifesting significant hypertension (Table 1).

Fig. 2.

Estimated cardiovascular disease (CVD) [peripheral vascular disease (PVD)] risk across animal groups used in the present study. Estimated PVD risk is determined based on the severity of obesity, impaired glycemic control, dyslipidemia, hypertension, systemic inflammation, and systemic oxidant stress in each animal group at the time of use (from data summarized in Table 1). SDR, Sprague-Dawley rats; LZR, lean Zucker rats; HSD, SDR on high-salt diet; HFD, SDR on high-fructose diet; RRM reduced renal mass-hypertensive SDR; SHR, spontaneously hypertensive rats; OZR, obese Zucker rats; DSS, Dahl salt-sensitive rats. See text for additional details.

Figure 3 presents the data describing the ex vivo reactivity of skeletal muscle resistance arterioles in response to endothelium-dependent dilation (acetylcholine), myogenic activation, or adrenergic constriction (phenylephrine). Dilator responses to acetylcholine, robust in SDR and LZR, were significantly attenuated in all of the conditions of elevated PVD risk, as evidenced by a significant reduction in the upper bound of the concentration-response relationship (Fig. 3A). Pretreatment with TEMPOL resulted in a significant increase in dilator responses in HFD, RRM, SHR, OZR, and DSS groups. In all cases, pretreatment with l-NAME abolished dilator responses to acetylcholine. Myogenic activation of skeletal muscle resistance arterioles (Fig. 3B), modest in SDR and LZR, was significantly increased in all of the models of elevated PVD risk, with the exception of the HSD animals, where pressure-induced constriction of arterioles was unaffected. TEMPOL and l-NAME had only mild effects on the slope of myogenic activation and generally did not reach statistical significance. Constrictor responses to phenylephrine were robust in all groups, with responses in OZR and DSS being the most vigorous. TEMPOL had modest effects to blunt constrictor responses in the conditions of elevated PVD risk, while pretreatment with l-NAME exhibited an effect in SDR and LZR only (Fig. 3C). During combined TEMPOL and l-NAME treatment, responses to acetylcholine, increased intraluminal pressure, and phenylephrine were nearly identical to those for l-NAME alone (data not shown).

Fig. 3.

Indexes of reactivity of ex vivo gracilis muscle resistance arterioles isolated from each group of animals used in the present study. Data are presented as the maximum bound of the acetylcholine (ACh) concentration-response relationship (A), the slope of the myogenic activation relationship, where increasing negativity reflects a more robust pressure-induced constriction (B), and the maximum bound of the phenylephrine (Phe) concentration-response relationship (C). Data (means ± SE) are presented under control conditions and after pretreatment of the isolated vessel with either the antioxidant TEMPOL or the NOS inhibitor nitro-l-arginine methyl ester (l-NAME). *P < 0.05 vs. responses in SDR or LZR. †P < 0.05 vs. responses in that group under control conditions. See text for details.

Figure 4 presents the reactivity of distal arterioles within in situ cremaster muscle of the different rat groups in response to increasing concentration of acetylcholine (Fig. 4A) or arachidonic acid (Fig. 4B). Dilator responses of distal arterioles to acetylcholine were significantly reduced with all levels of elevated PVD risk, and the severity of the elevated risk did not appear to impact the magnitude of the impaired response. Treatment of the cremaster muscle with TEMPOL generally increased dilator responses to acetylcholine, although this was not the case in HSD. Arachidonic acid-induced dilation demonstrated a more progressive impairment in reactivity, with increasing PVD risk across the animal groups (Fig. 4B), and also demonstrated an improved dilator response following pretreatment of the cremaster muscle with TEMPOL. Interestingly, treatment of the cremaster muscle with SQ-29548 (to block PGH₂/TxA₂ receptors) resulted in dilator responses that were very similar to those for TEMPOL pretreatment. In all cases, treatment with indomethacin severely attenuated vascular reactivity to increasing concentrations of arachidonic acid.

Fig. 4.

Dilator reactivity of in situ distal arterioles of the cremaster muscle within each group of animals used in the present study. Data are presented as the maximum bound of either the acetylcholine concentration-response relationship (A) or the arachidonic acid concentration-response relationship (B). Data (means ± SE) are presented under control conditions and after pretreatment of the isolated vessel with the antioxidant TEMPOL (both), the NOS inhibitor l-NAME (acetylcholine only), the PGH₂/TxA₂ receptor antagonist SQ-29548 (arachidonic acid only), or the cyclooxygenase inhibitor indomethacin (INDO). *P < 0.05 vs. responses in SDR or LZR. †P < 0.05 vs. responses in that group under control conditions. ‡P < 0.05 vs. responses in that group under TEMPOL-treated conditions. See text for details.

The bioavailability of key signaling molecules within arteries of the different groups of animals is summarized in Fig. 5. Vascular NO bioavailability, based on the slope of the methacholine vs. NO production relationship, was robust in healthy SDR and LZR but demonstrated a severe reduction in all models of PVD risk, regardless of the disease model (Fig. 5A). In all models except for HSD, pretreatment of vessel with TEMPOL improved NO bioavailability. Vascular production of PGI₂, estimated from its stable breakdown product 6-keto-PGF_1α, was also highest in SDR and LZR and manifested a gradual decline as the severity of PVD risk increased (Fig. 5B). In general, pretreatment with TEMPOL improved PGI₂ release, and incubation with indomethacin abolished it. In contrast, vascular production of TxA₂, lowest in arteries from SDR and LZR, rose with increasing PVD risk severity across the other six models (Fig. 5C). Pretreatment with TEMPOL reduced TxA₂ production, and incubation with indomethacin severely attenuated it.

Fig. 5.

Vascular bioavailability of nitric oxide [NO, A; estimated from slope of the NO production vs. methacholine (Met) concentration relationship], production of PGI₂ (B; estimated from 6-keto-PGF_1α), and TxA₂ (C; estimated from 11-dehydro-TxB₂) across animal groups used in the present study. *P < 0.05 vs. responses in SDR or LZR. †P < 0.05 vs. responses in that group under control conditions. ‡P < 0.05 vs. responses in that group under TEMPOL-treated conditions. See text for details.

Figure 6 presents the results of the in situ muscle perfusion experiments in the animal groups. In response to muscle contraction, the functional hyperemic response of the perfused skeletal muscle was robust in SDR and LZR in response to both 3-Hz and 5-Hz twitch contractions (Fig. 6A). With increasing severity of PVD risk score, the magnitude of the hyperemic response was progressively reduced, reaching statistical significance in OZR and DSS in response to 3 Hz and in SHR, OZR, and DSS in response to 5 Hz. Oxygen extraction across the skeletal muscle during contraction was generally stable in response to 3-Hz contraction but was reduced in SHR, OZR, and DSS at 3-Hz and OZR and DSS at 5-Hz twitch contraction (Fig. 6B). A similar pattern was identified in response to 5-Hz twitch contraction, although the differences were only significant in OZR and DSS compared with those in the healthy control animals. Muscle blood flow, oxygen extraction, and oxygen consumption (V̇o₂; Fig. 6C) demonstrated a gradual reduction from levels determined in SDR and LZR with increasing PVD risk severity, reaching statistical significance in OZR and DSS for both 3-Hz and 5-Hz contractions (and in SHR for 3 Hz only). Muscle performance, defined as the extent to which the maximum twitch tension was maintained after 3 min of either 3-Hz or 5-Hz stimulation across the animal groups, is summarized in Fig. 6D. In response to either 3-Hz or 5-Hz twitch contraction, the ability of the skeletal muscle to maintain contractile tension was progressively attenuated with increasing PVD risk score, achieving statistical significance in OZR and DSS at 3 Hz and in OZR, DSS, and SHR at 5 Hz.

Fig. 6.

Vascular responses and contractile performance of in situ skeletal muscle of animal groups in the present study in response to 3 min of muscle contraction at 3 or 5 Hz (isometric twitch). Data are presented for hyperemic responses to muscle contraction (A), oxygen extraction across the gastrocnemius muscle (B), oxygen consumption across the gastrocnemius muscle (C), and % of peak force development after 3 min of the contraction bout (D). *P < 0.05 vs. responses in SDR or LZR at that contraction frequency. See text for details.

In the in situ self-perfused cremaster preparation, data describing the changes in the arteriolar bifurcation distribution coefficient (γ) throughout the microvascular network are summarized in Fig. 7. At all levels of the microvascular network, the average γ during the observation window was closest to homogeneous sharing between daughter arterioles in healthy SDR and LZR. With increasing severity of PVD risk, γ generally increased in parallel with aggregate PVD risk, reaching maximum levels in OZR and DSS. However, it is apparent that the changes in aggregate γ with increased PVD risk may have been manifested first in distal arteriolar bifurcations (Fig. 7, C and D) and, although present, did not significantly impact perfusion distribution in the more proximal microcirculation with milder levels of PVD risk (Fig. 7, A and B).

Fig. 7.

Microvascular perfusion distribution (γ) at arteriolar bifurcations within in situ cremaster muscle. Data are presented as means ± SE for each animal group spanning 1A-2A arterioles (A), 2A-3A arterioles (B), 3A-4A arterioles (C), and 4A-5A arterioles (D). *P < 0.05 vs. SDR or LZR. See text for additional details.

With these data, and the basic simulation described above, the frequency distribution of blood perfusion at the terminal arteriolar level within the microvascular network across the different groups can be estimated (Fig. 8). As a result of the steady shifts in γ at successive arteriolar bifurcations with increasing PVD risk (Fig. 8, A–H), the distribution of perfusion at the precapillary level in the network moves from a more traditionally Gaussian distribution and assumes a distribution that, while still retaining characteristics of a lagged normal distribution, becomes increasingly broadly distributed and right-skewed (13), with a large number of “ischemic” regions/pathways at the expense of a small number of regions/pathways that experience extremely high or “hyperemic” perfusion.

Fig. 8.

Predicted perfusion distributions in the distal (precapillary) microcirculation of skeletal muscle of the animal groups within the present study. Frequency distributions are calculated based on an 8-bifurcation network using the microvascular perfusion distribution coefficients (γ) determined in the in situ cremaster muscle presented in Fig. 7. A–H present the distribution of perfusion across of 256 (2⁸) parallel arterioles under each experimental condition resulting from the simulation of a dichotomous branching network. Please see text for additional details.

The remaining presentation of data will focus on “proximal” arteriolar bifurcations (1A-2A) and “distal” arteriolar bifurcations (3A-4A). As the transition between bifurcation levels is gradual, this allows for a focus on key concepts while minimizing unnecessary duplication/repetition. The temporal behavior of proximal and distal arteriolar bifurcations along the microvascular network in the different groups of animals is summarized in Fig. 9. For proximal bifurcations (1A-2A), the cumulative changes in γ are presented in Fig. 9A, demonstrating no clear trend or obvious net movement in γ. However, when presented as the changes in γ over time summated as absolute differences, the progressive loss of temporal activity in γ at proximal bifurcations with increasing PVD risk is clearly evident (Fig. 9B). A similar pattern is evident in distal arteriolar bifurcations (3A-4A), where no clear trend is evident when the temporal changes to γ are simply summated (Fig. 9C), although when summated as absolute changes over time, a significant reduction in the activity of γ over time becomes clearly evident with increasing PVD risk score (Fig. 9D). Interestingly, the changes determined in smaller arteriolar bifurcations (Fig. 9D) did not exhibit the progressive decay with increasing PVD risk that was present in larger arteriolar bifurcations (Fig. 9B); rather, the changes to γ appeared to be very sensitive to any elevation of PVD risk and less susceptible to the degree of severity.

Fig. 9.

Data describing the cumulative changes in γ over the minute collection period in 1A-2A (A and B) and 3A-4A (C and D) arteriolar bifurcations in the animal groups of the present study. Cumulative changes in γ (presented as means ± SE) are summated either as sequential differences between successive time points (A and C) or as the sequential absolute differences between successive time points (B and D). *P < 0.05 vs. SDR or LZR. †P < 0.05 vs. HSD or HFD. ‡P < 0.05 vs. RRM or SHR. See text for additional details.

Figure 10 presents the attractors describing the spatio-temporal changes to γ in the proximal microcirculation (1A-2A arteriolar bifurcations) across the groups in the present study. Figure 10A presents data from the control “healthy” models, SDR and LZR, where the attractor occupied a large volume of space and was highly variable with time. In subsequent panels of Fig. 10, data for SDR and LZR have been “grayed” to facilitate comparisons to control models. With “mild” PVD risk, as shown in HSD and HFD rats in Fig. 10B, minimal changes were evident in the behavior of the attractors compared with that in control animals. As PVD risk become stronger (RRM and SHR; Fig. 10C), the central region of the attractor began to shift away from that in control animals and its range began to become increasingly constrained. With high PVD risk (OZR and DSS; Fig. 10D), this shift became more severe, with the shape of the attractor indicating increasingly heterogeneous perfusion distribution and a severely constrained flexibility.

Fig. 10.

Presentation of the attractor describing the overall spatial-temporal behavior of γ at 1A-2A arteriolar bifurcations across the animal groups of the present study. The attractors are presented as iterated maps, where the respective value for γ is presented at multiple successive time points (t) within that group. A presents the attractor for 1A-2A arteriolar bifurcations in SDR (blue) and LZR (dark green); these data are grayed in B–D to facilitate comparisons. B: attractor for 1A-2A arteriolar bifurcations in HSD (light green) and HFD (cyan). C: attractor for 1A-2A arteriolar bifurcations in RRM (brown) and SHR (dark green). D: attractor for 1A-2A arteriolar bifurcations in OZR (black) and DSS (purple). See text for additional details.

While a similar pattern with the change in the shape and location of the attractor was also present for the distal microcirculation (3A-4A arteriolar bifurcations; Fig. 11), some fundamental differences were revealed. Although responses in SDR and LZR were very comparable (Fig. 11A), the presence of mild PVD risk in HSD and HFD resulted in a clear reduction in the size of the attractor with a slight shift in the central region toward increasingly heterogeneous perfusion distribution (Fig. 11B). This effect was increased further with moderate PVD risk (RRM and SHR; Fig. 11C). In the setting of high PVD risk, the shift in the central region of the attractor was clearly evident, with a severe reduction in both its size and flexibility (Fig. 11D).

Fig. 11.

Presentation of the attractor describing the overall spatial-temporal behavior of γ at 3A-4A arteriolar bifurcations across the animal groups of the present study. The attractors are presented as iterated maps, where the respective value for γ is presented at multiple successive time points within that group. A presents the attractor for 3A-4A arteriolar bifurcations in SDR (blue) and LZR (dark green); these data are grayed in B–D to facilitate comparisons. B: attractor for 3A-4A arteriolar bifurcations in HSD (light green) and HFD (cyan). C: attractor for 3A-4A arteriolar bifurcations in RRM (brown) and SHR (dark green). D: attractor for 3A-4A arteriolar bifurcations in OZR (black) and DSS (purple). See text for additional details.

DISCUSSION

The ultimate challenge facing clinical outcomes and public health with regard to elevated PVD risk is the effect that a mismatch in perfusion:demand relationships has on tissue/organ function in the more acute time frame, leading to the establishment of chronic conditions of suboptimal control of perfusion across longer periods of time and thus impacting on patient quality of life and mortality. However, some of the questions that this immediately poses are as follows: What is the optimal marker of perfusion status within a tissue in terms of understanding the impact of elevated PVD risk? How effective is vascular reactivity at predicting outcomes? How are the mechanisms of vascular reactivity integrated to produce an outcome? How informative is bulk blood flow to an organ under different challenged conditions? Are there markers of perfusion control/status that provide greater insight into the state of the microvascular networks and how effectively they are behaving under a myriad of conditions? The purpose of the present study was to provide some insight into the multiscale behavior of the skeletal muscle microcirculation under conditions of progressively elevated PVD risk to begin to link indexes of microvascular dysfunction to muscle blood flow and performance outcomes.

At the level of the ex vivo skeletal muscle resistance arteriole, vascular reactivity was progressively altered with increasing PVD risk. As has been demonstrated previously for the experimental conditions of the present study, all of the conditions of elevated PVD risk were associated with an impaired endothelial function, as evidenced by a reduced upper bound of the acetylcholine concentration-response relationship (9, 17, 19, 24, 33, 34, 39, 49, 58), although the degree of impairment was similar in magnitude regardless of the severity of the PVD risk score. However, the contribution of oxidant stress to each condition of PVD risk was more variable, with some of the states of more severe PVD risk showing a robust role for oxidant stress in compromising endothelial function while others demonstrated a more modest or minimal contribution. In contrast, adrenergic reactivity was less affected until the models of PVD risk began to include obesity and hypertension. While there are multiple reports that alterations to adrenergic signaling are a major contributor to phenylephrine-induced (constrictor) responses (6, 10, 41, 52), there is also considerable evidence in the existing literature that the impairments to endothelial function result in the loss of a “buffering” of adrenergic responses, such that these are enhanced independent from any changes to adrenergic signaling per se (29, 42, 43). Myogenic activation tended to be mildly increased with conditions of elevated PVD risk that included hypertension, potentially for a protective effect on the downstream microcirculation (14, 59). For detailed recent reviews regarding vascular adaptations to obesity and insulin resistance, the reader is directed to References 46 and 49.

It is generally assumed that these kinds of alterations to vascular reactivity will serve to either elevate vascular resistance or blunt its appropriate reduction across levels of metabolic demand. However, the results from the present study bring this conceptual paradigm into question, as several of the conditions associated with significant alterations to vascular reactivity, extremely low NO bioavailability (HSD, HFD, RRM), and an elevation in TxA₂ production (HFD, RRM, SHR) were associated with minimal (if any) reductions to either skeletal muscle functional hyperemic responses or fatigue resistance with either moderate (3-Hz contractions) or high (5-Hz contractions) elevations in metabolic demand. As shown by the results of the present study, reductions in functional hyperemic responses or muscle fatigue resistance were only evident with considerably more severe conditions of elevated PVD risk. Although not assessed in the present study, these may have also been associated with a structural remodeling of the microvascular networks (e.g., capillary rarefaction; Refs. 23, 38). This may represent a well-maintained compensatory mechanism against PVD that begins to fail rapidly when an overburdening threshold is reached.

Under a setting of elevated PVD risk and the evolving ischemic condition, increasing O₂ extraction would have been expected as a compensatory mechanism (owing to increased mean resident time in exchange vessels) within the skeletal muscle. The results from the present study clearly suggest that this was not the case, as both O₂ extraction and V̇o₂ fell slightly during the imposed elevations in metabolic demand with increasing PVD risk. As there is no a priori reason to hypothesize that mechanisms of mass transport and exchange would be impaired under conditions of elevated PVD risk to reduce O₂ extraction, this suggests that the pathways of blood perfusion across skeletal muscle must be altered from that in a state of low PVD risk. However, it is important to note that that the potential for microvascular rarefaction is unclear and could represent a significant contributing factor to the altered blood gas exchange data (50).

Recent work using tracer washout analyses (4, 60) has suggested that the distribution of blood flow within skeletal muscle of OZR with full manifestation of the metabolic syndrome becomes increasingly heterogeneous. Furthermore, the multiscale validity of this interpretation was supported with higher-resolution techniques examining the distribution of erythrocytes and blood flow at individual arteriolar bifurcations within the in situ cremaster muscle of OZR (21). One interpretation of this previous work was that the establishment of conditions of increasingly heterogeneous perfusion distribution within the microcirculation that creates “high-flow” and “low-flow” pathways could result in the establishment of regions within the tissue that are characterized by excessively low perfusion while a smaller number of other areas of the tissue are associated with higher than normal perfusion. Although potentially problematic for the optimization of muscle fatigue resistance, the issue of greater global concern was the observation that the potential for compensation within the system was attenuated and that this environment, characterized by increasingly heterogeneous perfusion distribution, was also less flexible. The combination of these effects would result in the creation of a condition of increasing blood flow maldistribution that is also increasingly resistant to modification to adapt to any imposed challenge. These accumulated disruptions in microvascular compensatory mechanism against PVD likely lead to the overt macrovascular failure that has significant clinical morbidity and mortality effects.

The results from the present study begin to place this new conceptual understanding of microvasculopathy with elevated PVD risk into a more realistic context. Analyses of the perfusion distribution coefficient (γ) at successive levels within the microvascular networks provide compelling evidence that the severity of the alterations to γ were roughly correlated with the extent to which PVD risk was elevated. An interesting element of this finding is that the alterations to γ appear to develop in the smaller microvessels and bifurcations first (with the lowest level of PVD risk) and progress toward the larger microvessels and bifurcations with increasing PVD risk. This clearly suggests that the smaller microvessels may be more sensitive to the compromised environment associated with PVD risk but the decay in function at this level may not be sufficient to result in a reduction in the hyperemic responses and fatigue resistance of skeletal muscle under conditions of elevated metabolic demand.

While the accumulated effect of the changes to γ with elevated PVD risk may result in a progressive increasingly heterogeneous distribution of perfusion at the precapillary level (Fig. 8), of greater concern with regard to integrated microvascular function and failure is the reduction in the activity (i.e., changes in γ over time) at the successive arteriolar bifurcations. While the general trend toward a reduction in temporal activity throughout the microcirculation with increasing PVD risk was clearly evident, a fundamental difference that exacerbates the spatial distribution concern discussed above was also present. In the larger microvessels and bifurcations, there was a clear correlation between the degree to which temporal activity was blunted and the severity of the PVD risk. However, in the smaller microvessels and bifurcations this was not the case, as temporal activity was significantly attenuated with even mild PVD risk and was not further impacted despite more severe levels of risk.

This concept of an early failure of the microcirculation with mild PVD risk and a progressive failure of the proximal microcirculation with more moderate-high levels of risk is most clearly demonstrated in the attractor data (Figs. 10 and 11). These data clearly demonstrate the incremental impact of elevated PVD risk on the shifted position and reduced variability in γ and reveal that vascular dysfunction (at least in terms of control of tissue perfusion) may begin in the smaller microvessels and bifurcations with only mild elevations in risk and progress in a retrograde fashion toward the proximal microcirculation as PVD risk levels increase.

It has become commonly accepted that NO bioavailability represents a critical contributor to the regulation of vascular tone under normal physiological conditions, and that the loss of this pathway represents a major element in vasculopathies and altered mass transport and exchange during challenged conditions (27, 35, 47, 48, 54). In this regard, the results from the present study may provide new insight into these relationships. While all of the conditions of elevated PVD risk were associated with comparable reductions in vascular NO bioavailability, those conditions of mild elevations of PVD risk demonstrated no changes to blood flow, blood gas handling, and muscle performance compared with the normal, healthy conditions. However, under these conditions, the reductions in NO bioavailability were associated with a profound loss of attractor flexibility in the distal microcirculation. As such, it may be reasonable to speculate that the loss of NO bioavailability with even mild elevations in PVD risk may reflect the genesis of a condition of a compromised “buffer” against systemic challenges. As such, there may be no overt physiological phenotype associated with this loss to NO bioavailability itself. Rather, the environment may become sufficiently compromised that challenges that are not normally sufficient to impair physiological outcomes may become more damaging owing to the loss of NO bioavailability as a compensatory mechanism.

As PVD risk gets progressively worse, and oxidant stress and inflammation become exacerbated, this increasingly prooxidant and proinflammatory environment becomes sufficiently large to shift arachidonic acid metabolism from its normal pathways to ones characterized by elevated production of TxA₂. While it is certainly likely that this “balance” is far more complex than simply PGI₂ vs. TxA₂, and may involve substances that are associated with other pathways of arachidonic acid metabolism (e.g., lipoxygenase, cytochrome P-450 pathways, etc.), our conjecture will remain somewhat limited at this time. A shift in the balance toward TxA₂ genesis clearly contributes to the reduction in the attractor in the distal microcirculation and also contributes to the reduced variability of the attractor for the proximal microcirculation. Essentially, this may act as a “dampener” of entire microvascular networks and further constraint on the reactivity of the entire network rather than the distal microcirculation (as appears to be the case for NO bioavailability). Evidence from other research groups has provided support for this conjecture, demonstrating the impact of elevated TxA₂ activity to blunt agonist-induced and functional hyperemic responses within blood-perfused in situ microvascular networks (32, 61, 62).

A similar outcome on reactivity throughout the network may follow from the graded increase in myogenic activation with elevated PVD risk. While changes to myogenic activation were modest until the development of hypertension (where the increased pressure-induced constriction was enhanced), the increased sensitivity of myogenic activation under conditions of PVD risk that include uncontrolled hypertension would impose an additional level of constrictor tone within the proximal microcirculation. This chronic increase in the level of myogenic tone may act to further dampen the flexibility of the system and constrain its ability to adapt to stressors and challenges.

The contribution of adrenergic constriction to the changes in the position and shape of the attractors may be somewhat more difficult to assess. While certainly some of these effects are the result of the changes to endothelial function impacting the constrictor responses (29, 42, 43), there is previous evidence to support increases in adrenergic activity (6), receptor sensitivity (52), or downstream signaling gain (41) as potential contributors to increased constrictor responses in many of these models. While this would potentially reduce attractor variability further and would represent an additional, and possibly extremely robust, depressor of arteriolar function and flexibility in the control of tissue perfusion, it is unclear that a simple “increase in adrenergic constrictor reactivity” accurately represents the prevailing condition in conditions of elevated PVD risk that are associated with obesity and hypertension (28, 30, 37). If this was a widespread phenomenon, it would result in either enormous elevations in perfusion pressure or dramatic reductions in blood flow. This is not clearly evident in several of the models (e.g., OZR, SHR), where the elevations in blood pressure, while considerable, are not associated with profound reductions in blood flow. Clearly, while there is a central role for adrenergic constriction in contributing to the changes to the attractor (especially at proximal arteriolar bifurcations), the actual role for adrenergic responses appears to be more complicated than a simple shift in responses, and this will require additional targeted investigation that is beyond the scope of the present study.

Given that the results of the present study span eight animal models of human health and elevated PVD risk and employ multiple different protocols and procedures to establish multiscale validity, a degree of caution should be exercised when interpreting the data and extrapolating them to human and/or clinical relevance. While these are all animal models for common patterns of elevated PVD risk in humans, direct translational relevance may be limited by species and strain differences. Furthermore, the present study uses data from gracilis, cremaster, and gastrocnemius muscles and should be interpreted in that context, rather than as broadly representative of skeletal muscle as a whole.

The functional implications for the changes to the shape and position of the attractors with elevated PVD risk may be critical to our integrated understanding of the effects of these challenges on tissue/organ function. In general, the shape of the attractor provides us with an insight into the degree to which a system, in this case the microvascular network, can change its state to respond to imposed challenges and stressors. When the attractor is large and diverse, a greater number of potential conformational states are available for the arteriolar bifurcation, providing it with a greater degree of flexibility to adapt to altered conditions (either physiological or pathological) without resulting in a microcirculation-dependent loss of tissue/organ function. However, as PVD risk grows, the shape and position of the attractor change to encourage not only an increasingly heterogeneous distribution of blood flow within the microvascular networks (the change in γ) but also a progressive reduction in the number of available conformational states that can be assumed throughout the network (the temporal activity) before tissue/organ function begins to fail. As such, a given level of challenge (e.g., increased metabolic demand from muscle contraction) that can be easily accommodated by the microvascular networks under conditions of health, or even mild elevations in PVD risk, cannot be accommodated with further elevations in PVD risk, and this is associated with the traditional “ischemic” markers associated with this condition. Interestingly, the results from the present study suggest that “blood flow” may not be a particularly informative marker of PVD risk and outcomes of PVD below a certain easily determined threshold. Rather, the spatial and temporal behavior of intraorgan perfusion distribution may be a far more relevant marker in terms of understanding PVD risk and tissue/organ outcomes. Determination of not only how the alterations to the different mechanisms for the regulation of vascular resistance are manifested (i.e., vascular dilator/constrictor reactivity), and how these might combined with alterations to the structure and function of the endothelial glycocalyx as determinants of microvascular hematocrit (16, 55, 56), appears to be a next logical step toward an integrated understanding of the impact of PVD risk on microvascular function. The extent to which the changing position and shape of the attractor represent a broadly applicable concept that can contribute to the impact of PVD risk on other critical tissues/organs and some of the major health outcomes associated with these conditions (e.g., myocardium for angina, cerebral structures for stroke/cognitive deficits, kidney for volume and pressure regulation) may represent compelling areas for future investigation and intervention.

GRANTS

This study was supported by the American Heart Association (IRG 14330015, PRE 16850005, EIA 0740129N) and the National Institutes of Health (1P20 GM-109098, RR-2865AR; P20 RR-016477, R01 HL-065289, R01 HL-37374, R01 DK-64668).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: J.C.F., J.T.B., S.J.F., I.M.O., P.D.C., A.G.G., P.A.S., R.W.B., and J.H.L. conception and design of research; J.C.F., J.T.B., and A.G.G. performed experiments; J.C.F., J.T.B., and S.J.F. analyzed data; J.C.F., J.T.B., S.J.F., I.M.O., P.D.C., L.E.T., A.C.d., C.D.S., A.G.G., P.A.S., S.D.B., R.W.B., and J.H.L. interpreted results of experiments; J.C.F. prepared figures; J.C.F., J.T.B., S.J.F., P.D.C., L.E.T., A.C.d., C.D.S., and J.H.L. drafted manuscript; J.C.F., J.T.B., S.J.F., I.M.O., P.D.C., L.E.T., A.C.d., C.D.S., A.G.G., P.A.S., S.D.B., R.W.B., and J.H.L. edited and revised manuscript; J.C.F., J.T.B., S.J.F., I.M.O., P.D.C., L.E.T., A.C.d., C.D.S., A.G.G., P.A.S., S.D.B., R.W.B., and J.H.L. approved final version of manuscript.