Impaired diaphragm resistance vessel vasodilation with prolonged mechanical ventilation

Abstract

Mechanical ventilation (MV) is a life-saving intervention, yet with prolonged MV (i.e., ≥6 h) there are time-dependent reductions in diaphragm blood flow and an impaired hyperemic response of unknown origin. Female Sprague-Dawley rats (4–8 mo, n = 118) were randomized into two groups; spontaneous breathing (SB) and 6-h (prolonged) MV. After MV or SB, vasodilation (flow-induced, endothelium-dependent and -independent agonists) and constriction (myogenic and α-adrenergic) responses were measured in first-order (1A) diaphragm resistance arterioles in vitro, and endothelial nitric oxide synthase (eNOS) mRNA expression was quantified. Following prolonged MV, there was a significant reduction in diaphragm arteriolar flow-induced (SB, 34.7 ± 3.8% vs. MV, 22.6 ± 2.0%; P ≤ 0.05), endothelium-dependent (via acetylcholine; SB, 64.3  ± 2.1% vs. MV, 36.4 ± 2.3%; P ≤ 0.05) and -independent (via sodium nitroprusside; SB, 65.0 ± 3.1% vs. MV, 46.0 ± 4.6%; P ≤ 0.05) vasodilation. Compared with SB, there was reduced eNOS mRNA expression (P ≤ 0.05). Prolonged MV diminished phenylephrine-induced vasoconstriction (SB, 37.3 ± 6.7% vs. MV, 19.0 ± 1.9%; P ≤ 0.05) but did not alter myogenic or passive pressure responses. The severe reductions in diaphragmatic blood flow at rest and during contractions, with prolonged MV, are associated with diaphragm vascular dysfunction which occurs through both endothelium-dependent and endothelium-independent mechanisms.

NEW & NOTEWORTHY Following prolonged mechanical ventilation, vascular alterations occur through both endothelium-dependent and -independent pathways. This is the first study, to our knowledge, demonstrating that diaphragm arteriolar dysfunction occurs consequent to prolonged mechanical ventilation and likely contributes to the severe reductions in diaphragmatic blood flow and weaning difficulties.

INTRODUCTION

Mechanical ventilation (MV) is a life-saving intervention to sustain pulmonary gas exchange in patients who are incapable of maintaining sufficient alveolar ventilation (e.g., patients with respiratory failure, spinal cord injury, and heart valve replacement), with the frequency of MV use increasing by >2% annually (30). Prolonged MV in humans (≥18 h) and ≥6 h in the preclinical animal model (15, 18, 35, 37, 39) induces diaphragm atrophy, contractile dysfunction, oxidative stress, and mitochondrial dysfunction. Collectively, these perturbations, occurring primarily with prolonged MV, are known as ventilator-induced diaphragmatic dysfunction (VIDD) (45). The development of VIDD is believed to be a key culprit in the inability to wean patients from the ventilator and is a major clinical problem as ~30% of patients experience weaning difficulties (6, 14), resulting in additional time on the ventilator and increased patient mortality (14, 15).

The precise physiological mechanisms responsible for VIDD, whilst unclear, are likely to be multifaceted. Previous studies have indicated increased reactive oxygen species (ROS) production, mitochondrial damage, skeletal muscle atrophy, and muscle fiber remodeling as potential contributors to diaphragm weakness and VIDD (15, 18, 35, 37, 39). One aspect of VIDD that has not been explored is the potentially altered vasomotor control of the resistance vasculature within the diaphragm (i.e., arterioles). In this regard, it is established that during acute MV (e.g., 30 min) blood flow to the diaphragm is reduced (17, 36, 44). Recently, our group has expanded these observations, demonstrating an additional reduction in diaphragmatic blood flow when MV is extended from 30 min to 6 h (10). The mechanistic bases for this time-dependent reduction in diaphragmatic blood flow with MV are unknown. The inability to augment diaphragm blood flow with contractions after prolonged MV (10) suggests a reduced dilatory capacity and/or enhanced contractile responses of diaphragm resistance vessels occurring with prolonged MV. Importantly, this time-dependent reduction in blood flow with inactivity is not observed in other highly oxidative muscles (e.g., the soleus and red portion of the gastrocnemius muscle) (10), suggesting that the diaphragm vasculature may be particularly vulnerable to inactivity-induced rapid vascular dysfunction. Currently, there are no data regarding the impact of prolonged MV on vasomotor control of the resistance vessels within the diaphragm.

Vasomotor control demonstrates considerable temporal plasticity. For example, Woodman and colleagues (46) found that increasing intraluminal flow in skeletal muscle feed arteries of senescent rats upregulated endothelium-dependent vasodilation and endothelial nitric oxide synthase (eNOS) mRNA expression in as little as 4 h. However, there is a paucity of data regarding the ramifications of reductions in blood flow on vasomotor function in chronically active muscle such as the diaphragm. Whether functional downregulation is evident from reduced shear stress associated with inactivity and alters vascular control (e.g., blunted endothelium-dependent vasodilation) and/or pathways associated with nitric oxide production (e.g., eNOS) is unknown. There is evidence, from long-term (i.e., days/weeks) disuse, for a downregulation of vasodilatory pathways in skeletal muscle (11).

The overall objective of this study was to investigate the effects of prolonged MV (i.e., 6 h), using an established preclinical animal model, on diaphragm resistance arteriole function. In resistance vessels from the diaphragm of animals subjected to prolonged MV vs. spontaneous breathing, we tested the hypotheses that there will be 1) reduced endothelium-dependent and preserved endothelium-independent vasodilation, 2) an increased contractile response to both myogenic and α-adrenergic stimuli, and 3) no differences in passive pressure diameter responses. Results from these studies will provide potential mechanisms responsible for the severely diminished diaphragmatic hyperemic response with contractions after prolonged MV as well as pathways for intervention to address problematic weaning.

METHODS

Animals

Female Sprague-Dawley rats (4–8 mo old, ~350 g) were obtained from Charles River Laboratories (Boston, MA) for this investigation. Data were collected from either spontaneously breathing (SB; n = 64) animals or those successfully subjected (i.e., maintained arterial Po₂, Pco₂, and pH within normal ranges throughout the protocol, as described below) to prolonged (6 h) mechanical ventilation (MV; n = 54). In each animal, two first-order (1A) arterioles were used for the isolated vessel experiments, and the respective responses were averaged. Acetylcholine [ACh; ACh+N^G-nitro-l-arginine methyl ester (l-NAME), ACh+l-NAME and indomethacin] and sodium nitroprusside (SNP) dose-response experiments were performed in the same vessels from each group. Flow-mediated, eNOS mRNA quantification, active myogenic and passive pressure, and norepinephrine (NE) and phenylephrine (PE) dose-response experiments were performed independently in vessels from each group. The Sprague-Dawley rat was chosen due to the similar properties (e.g., anatomic and physiological) of its diaphragm to the human diaphragm (26, 27, 34). All procedures were approved by the Kansas State University and University of Florida Institutional Animal Care and Use Committees and complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Upon arrival, animals were housed and maintained in a temperature-controlled (23 ± 2°C) room with a 12:12-h light-dark cycle with water and rat chow provided ad libitum.

MV

All surgical procedures were performed using aseptic techniques. Animals in the MV groups were anesthetized with pentobarbital sodium (50 mg/kg ip), tracheostomized, and connected to a volume-cycled ventilator (Harvard Apparatus, Cambridge, MA or Kent Scientific PhysioSuite, Torrington, CT). A catheter was implanted in the carotid artery for measurement of blood pressure and periodic blood sampling (every 3 h) for analyses of blood gases (GemPremier 3000, Instrumentation Laboratory, Bedford, MA). Arterial O₂ saturation was monitored continuously by using a Mouse OX (Asbury Park, PA) placed around the rat’s foot. Expired Pco₂ was measured continuously using a microcapnograph (Model Columbia Instruments, Columbus, Ohio or Kent Scientific CapnoScan, Torrington, CT) and maintained at ≤30 mmHg to minimize carotid chemoreflex stimuli and render the diaphragm inactive. Arterial Po₂ and pH were maintained within the normal range (70–110 mmHg, and 7.35–7.45, respectively) by minor adjustments to minute volume. A catheter was placed in the jugular vein for infusion of pentobarbital sodium (10 mg·kg⁻¹·h⁻¹) and fluids, when necessary. Body temperature was monitored (via rectal thermometer) and maintained at 37 ± 1°C by use of a recirculating heating blanket. Body fluid homeostasis was maintained by administering an electrolyte solution [i.v., ~2 ml·kg⁻¹·h⁻¹; 0.9% saline 5.0 pH, Na⁺ (154 mM), Cl⁻ (154 mM)]. Operative care during the MV period included expressing the bladder, removal of airway mucus, lubricating the eyes, and rotating the animal. The ventilator was maintained at an average breathing frequency of 70 ± 10 breaths/min and tidal volume of 6–7 ml/kg body weight.

Isolated Microvessel Technique

To determine vasomotor control, resistance arterioles (<200-μm intraluminal diameter) were isolated from the medial costal diaphragm and studied in vitro to remove potentially confounding metabolic, humoral, and neural influences. Resistance arterioles were harvested from diaphragms of SB or prolonged MV animals. With the aid of a dissecting microscope (Olympus SVH10), 1A arterioles from the medial costal diaphragm muscle were isolated and removed from the surrounding muscle tissue as previously described (1, 2, 28, 29, 41). The arterioles (length, 0.5–1.0 mm; inner diameter, 90–175 μm) were transferred to a poly(methyl methacrylate) chamber containing albumin-supplemented physiological saline solution (PSS) equilibrated with room air. Each end of the arteriole was cannulated with glass micropipettes and secured with nylon suture (11–0, ophthalmic suture, Alcon Laboratories). Following cannulation, the microvessel chamber was transferred to the stage of an inverted microscope (Olympus IX70) equipped with a video camera (Panasonic BP310), video caliper (Colorado Video 307A), and data-acquisition system (PowerLab) for online recording of intraluminal diameter. Arterioles were initially pressurized to 90 cmH₂O with two independent hydrostatic pressure reservoirs. Leaks were detected by pressurizing the vessel and then closing the valves to the reservoirs and verifying that intraluminal diameter remained constant. Arterioles that exhibited leaks were discarded. Arterioles that were free from leaks were warmed to 37°C and allowed to develop initial spontaneous tone during a 30- to 60-min equilibration period. Endothelium-dependent (i.e., flow-induced and exposure to acetylcholine) and -independent (SNP), myogenic (active and passive), and α-adrenergic responses (via NE and PE) were assessed in the isolated vessels, as described below.

Vasodilator Reponses

Flow-mediated vasodilation.

After a steady level of spontaneous tone was achieved, diaphragm arterioles were exposed to graded increases in intraluminal flow. This was accomplished by altering the heights of the independent fluid reservoirs in equal and opposite directions so that a pressure difference was created across the vessel without altering mean intraluminal pressure, as previously described (29). Diameter measurements were then determined in response to incremental pressure differences of 4, 10, 20, 40, and 60 cmH₂O. These step changes in pressure elicit increases in intraluminal flow corresponding to physiologically relevant flow (29) and shear stress (5) rates.

ACh-mediated vasodilation.

Upon display of a steady level of spontaneous tone, vasodilatory responses to ACh (1 × 10⁻⁹ to 1 × 10⁻⁴ M), which mediates smooth muscle relaxation indirectly by binding to endothelial cell M₂ receptors and stimulates the release of endothelium-derived relaxing factor(s) (EDRF), were tested. EDRFs typically released from vascular endothelial cells are nitric oxide (NO) and prostacyclin (PGI₂), and to a lesser extent endothelium-derived hyperpolarizing factor (EDHF). Therefore, the contribution of the NOS and cyclooxygenase (COX) signaling pathways to ACh-induced vasodilation in the diaphragm resistance arterioles was determined following a 20-min incubation with 10⁻⁵ M of the NOS inhibitor l-NAME. Thereafter, the same vessel underwent a 20-min incubation period with 10⁻⁵ M l-NAME plus 10⁻⁵ M indomethacin.

SNP-mediated vasodilation.

To determine endothelium-independent function of medial costal diaphragm resistance arterioles, dose-responses to the exogenous NO donor (10⁻⁹ to 10⁻⁴ M), a potent vasodilatory agent that acts directly on smooth muscle through NO-cGMP activation to elicit vasodilation, were assessed.

Vasoconstrictor Responses

Active and passive myogenic responses.

Active myogenic responses were assessed by allowing the vessels to equilibrate at 37°C and 90 cmH₂O for 60 min, which allowed for the development of spontaneous tone. After equilibration, intraluminal pressure was increased in 10-cmH₂O increments from 0 to 140 cmH₂O. The diameter was recorded for 3 min at each pressure step, and these pressure changes occurred in the absence of flow. Thereafter, to assess the passive response, the vessels were incubated in calcium-free PSS for 60 min (bath changed every 20 min) at 37°C and 90 cmH₂O. Following equilibration, intraluminal pressure was again increased in 10-cmH₂O increments from 0 to 140 cmH₂O, and the diameter was recorded for 3 min at each pressure step in the absence of flow.

α-Adrenergic responses.

Vessels used for α-adrenergic dose-responses were equilibrated in PSS at 37°C at 90 cmH₂O for 60 min and rinsed every 20 min with PSS. Thereafter, contractile responses to cumulative doses (10⁻⁹ to 10⁻⁴ M) at 3-min intervals of NE and PE were determined, and the intraluminal diameter was measured at the end of each 3-min interval.

After completion of the final concentration-response or pressure-response, vessels were incubated with warm solution (pH = 7.4), free of calcium, for 60 min and were rinsed every 20 min with calcium-free PSS.

Following a 60-min normalization period, vessels were exposed to one 40-μl (10⁻⁴M) dose of SNP, and maximal diameters were recorded.

mRNA Analysis

Medial costal diaphragm resistance arterioles were immediately excised after euthanasia (see above), snap frozen, and stored at −80°C. eNOS mRNA expression was quantified with preformulated TaqMan primers (Applied Biosystems) for the following genes: GAPDH (Rn01775763_g1); 18s ribosomal RNA (Rn03928990_g1); and eNOS (Rn02132634_s1). Arterioles were pulverized in lysis buffer, and total RNA was extracted using an RNAqueous isolation kit (Ambion). Total mRNA (1 µg) was then reverse transcribed using the High-Capacity cDNA Archive kit (Applied Biosystems). Real-time PCR was performed as previously described (41). Briefly, relative quantitative real-time RT-PCR was performed using the TaqMan Fast Advanced Master Mix, and reactions were performed in duplicate using 96-well optical plates on an ABI Prism 7900HT fast real-time PCR system. Candidates for housekeeping genes, GAPDH and 18s, were tested for stability over various experimental conditions (data not shown). 18s was used as the stable housekeeping gene (endogenous control) to normalize the samples from the resistance arterioles isolated from the medial costal diaphragm. The expression levels were calculated in each experimental group in triplicate, with a minimum of two independent experiments. Relative quantitation was done using the ΔΔCT method, where CT is the cycle threshold, and all untreated samples were normalized to 1.

Data Analysis

Maximal diameter, body weight, and mRNA expression were analyzed with one-way ANOVA. Vasomotor responses were evaluated by two-way ANOVA to detect main effects between (SB and MV) or within (flow rate, pressure, inhibitor treatment, etc.) groups. Post hoc analyses were performed using Tukey’s test. All data are presented as means ± SE. Significance was established at P ≤ 0.05. Vasodilatory responses were recorded as actual diameters and represented as a percentage of maximal relaxation. Relaxation (%) of diaphragm arterioles was calculated according to the formula (29):

$$\text{Relaxation}\left(\%\right)=\left(D_{\text{s}}–D_{\text{b}}\right)/\left(D_{\text{m}}–D_{\text{b}}\right)$$

where D_m is the maximal inner diameter recorded at 90 cmH₂O in calcium-free PSS, D_s is the steady-state inner diameter recorded after each addition of the drug or increase in flow, and D_b is the initial baseline inner diameter recorded immediately before the first addition of the vasodilatory agent or initiation of flow. Spontaneous tone development and responses to NE and PE were expressed as the percent constriction relative to maximal diameter and was calculated using the formula (28):

$$\text{Spontaneous\hspace{0.17em}tone}\left(\%\right)=\left[\left({\text{ID}}_{\text{max}}–{\text{ID}}_{\text{b}}\right)/{\text{ID}}_{\text{max}}\right]\times 100$$

where ID_max represents the maximal inner diameter, and ID_b represents the baseline inner diameter following the 60-min equilibration period. Active myogenic diameter responses to changes in intraluminal pressure from 0 to 140 cmH₂O were normalized according to (28):

$$\text{Normalized\hspace{0.17em}diameter}={\text{(ID}}_{\text{s}}{\text{/ID}}_{\text{max}}\text{)}$$

where ID_s is the steady-state inner diameter measured after each incremental 10-cmH₂O pressure change. Vasoconstrictor responses to NE and PE were expressed as the percent change from baseline diameter based on the formula (28):

$$\text{Constriction}\text{(\%)}=\left[\left({\text{ID}}_{\text{b}}–{\text{ID}}_{\text{s}}\right){\text{/ID}}_{\text{b}}\right]\times \text{100}$$

where ID_b is the initial baseline diameter recorded immediately before the vasoconstrictor agonist addition, and ID_s represents the steady-state diameter recorded following each dose of the drug.

RESULTS

MV Reduced NO-Mediated Vasodilation

Spontaneous tone, maximal diameter, wall thickness, and wall-to-lumen ratio did not differ between groups (Table 1). Following prolonged MV, there was a 40% reduction in flow-induced vasodilation in diaphragm arterioles (Fig. 1; Table 1) and a significantly (~45%) diminished endothelium-dependent vasodilation to ACh (Fig. 2A; Table 1). In the presence of l-NAME, the response to ACh was significantly reduced in both groups (Fig. 2A, Table 1), although the reduction was comparatively greater in the SB vs. MV group (i.e., with l-NAME the dilation to ACh was reduced ~55% in the SB vs. ~30% in the MV group). Differences between groups to ACh were abolished in the presence of l-NAME (Fig. 2A). Compared with l-NAME alone, combined NOS and COX blockade with l-NAME and indomethacin, respectively, further reduced responses to ACh (Fig. 2B; Table 1) with no significant differences between groups. Flow-induced dilation (Max %) was reduced from 34.7 ±  3.8 to 22.6 ± 2.0% following prolonged MV (P ≤ 0.05; Table 1). Maximal responses to ACh between SB (63.3 ± 2.1%) and MV (36.4 ± 2.3%) vessels were significantly different (P ≤ 0.05; Table 1). Compared with SB, prolonged MV altered the sensitivity (EC₅₀) to ACh and ACh+l-NAME but not in the presence of l-NAME+indomethacin (Table 1). Diaphragm arteriolar eNOS mRNA expression was significantly reduced in the MV vs. SB group (Fig. 2C).

Table 1.

Diaphragm resistance vessel spontaneous tone development, vessel characteristics, and maximal responses from spontaneous breathing and 6-h mechanically ventilated rats

Measurement	SB (n = 64)	6-h MV (n = 54)
Vessel characteristics
Spontaneous tone, %	14.5 ± 1.9	11.4 ± 1.5
Maximal diameter, µm	160 ± 12	171 ± 11
Wall thickness, µm	22 ± 2	23 ± 2
Wall-to-lumen ratio, %	7.2 ± 0.9	7.4 ± 0.9
Maximal responses
Flow, Max %	34.7 ± 3.8 (n = 7)	22.6 ± 2.0 (n = 7)*
ACh, Max %	64.3 ± 2.1 (n = 22)	36.4 ± 2.3 (n = 20)
ACh+l-NAME, Max %	28.1 ± 1.2 (n = 22)#	25.0 ± 2.4 (n = 20)+
ACh+l-NAME+Indo, Max %	14.3 ± 1.5 (n = 22)#†	15.4 ± 1.1 (n = 20)+‡
SNP, Max %	65.0 ± 3.1 (n = 22)	46.0 ± 4.6 (n = 20)*
NE, Max %	33.2 ± 5.9 (n = 9)	26.1 ± 3.4 (n = 6)
PE, Max %	37.3 ± 6.7 (n = 9)	19.0 ± 1.9 (n = 6)*
EC50 [Log M] (10−7)
ACh	1.1 ± 0.4 (n = 22)	6.0 ± 2.1 (n = 20)*
ACh+l-NAME	2.9 ± 1.5 (n = 22)	36.0 ± 19.2 (n = 20)*
ACh+l-NAME+Indo	7.8 ± 3.7 (n = 22)	23.1 ± 21.3 (n = 20)
SNP	5.8 ± 3.7 (n = 22)	7.5 ± 3.3 (n = 20)
NE	5.3 ± 1.9 (n = 9)	43.0 ± 17.8 (n = 6)*
PE	23.0 ± 7.2 (n = 9)	93.0 ± 11.1 (n = 6)*

Values are means ± SE; n = vessels studied in vitro. SB, spontaneous breathing; MV, mechanical ventilation; ACh, acetylcholine; l-NAME, N^G-nitro-l-arginine methyl ester; Indo, indomethacin; SNP, sodium nitroprusside; NE, norepinephrine; PE, phenylephrine; EC₅₀, concentration of a drug that elicits a half-maximal response.

^*Significant (P ≤ 0.05) difference vs. SB.

⁺Significant (P ≤ 0.05) difference vs. 6-h MV ACh.

^‡Significant (P ≤ 0.05) difference vs. 6-h MV+l-NAME.

^#Significant (P ≤ 0.05) difference vs. SB ACh.

^†Significant (P ≤ 0.05) difference vs. SB ACh+l-NAME.

Fig. 1.

Flow-mediated vasodilation in diaphragm arterioles during spontaneous breathing (SB; n = 7) and after 6-h mechanical ventilation (MV; n = 7). *P ≤ 0.05 vs. SB.

Fig. 2.

A: dose-responses to the endothelium-dependent vasodilator acetylcholine (ACh) in the absence and presence of the endothelial nitric oxide (NO) synthase inhibitor N^G-nitro-l-arginine methyl ester (l-NAME). B: dose-responses to the endothelial-dependent vasodilator acetylcholine (ACh) in the absence and presence of the endothelial NO synthase inhibitor l-NAME+cyclooxygenase (COX) inhibitor indomethacin (INDO). C: endothelial nitric oxide synthase (eNOS) mRNA expression in diaphragm arterioles during SB and 6-h MV. D: dose-responses to the endothelium-independent vasodilator sodium nitroprusside (SNP) in diaphragm arterioles. Sample sizes for A, B, and D: SB (n = 22) and 6-h MV (n = 20). Sample size for eNOS mRNA expression in C: SB (n = 15) and 6-h MV (n = 13). *P ≤ 0.05 vs. SB ACh. +P ≤ 0.05 vs. 6-h MV ACh. #P ≤ 0.05 vs. SB ACh.

Following 6-h MV, there was a blunted endothelium-independent dilation to the exogenous NO donor SNP compared with responses of diaphragm arterioles from the SB group (Fig. 2D). In response to SNP, MV induced a ~20% reduction in maximal relaxation compared with SB (65.0 ± 3.1 vs. 46.0 ± 4.6%; P ≤ 0.05). Compared with SB, MV did not alter the EC₅₀ to SNP (Table 1). In the SB group, there were no differences in the relaxation (%) in response to ACh and SNP (Fig. 3A). There was a significant difference in relaxation (%) at the 10⁻⁵ (P ≤ 0.05) and 10⁻⁴ M (P ≤ 0.05) doses between ACh and SNP following 6-h MV (Fig. 3B).

Fig. 3.

ACh and SNP dose-responses in diaphragm arterioles from the SB group (n = 7; A) and from the 6-h MV group (n = 7; B). *P ≤ 0.05 vs. SNP dose-response.

Myogenic Vasoconstriction, Passive Pressure-Diameter Responses, and Contractile Responses to α-Adrenergic Agonists

There were no differences observed in the active myogenic response between groups (Fig. 4A) nor passive pressure-diameter responses between groups at all pressures (Fig. 4B). Vasoconstrictor responses to NE were not different between SB and MV (Fig. 5A; Table 1). However, with PE, maximal constriction (Max %) was diminished by 20% (Fig. 5B; Table 1) in vessels from the MV group (SB, 37.3 ± 6.7% vs. MV, 19.0 ± 1.9%; P ≤ 0.05). Compared with SB, EC₅₀ was altered to both NE and PE with prolonged MV (P ≤ 0.05; Table 1).

Fig. 4.

Active myogenic vasoconstriction (A) and passive pressure responsiveness (B) in diaphragm arterioles during SB (n = 11) and 6-h MV (n = 8). Active responses are presented in normalized diameter (see methods) to account for differences in initial potential differences in initial tone. Passive responses are presented as actual diameters. No change.

Fig. 5.

Dose-responses to nonselective α-adrenergic agonist norepinephrine (NE; A) and dose-responses to α₁-specific agonist phenylephrine (PE; B) during SB (n = 9) and 6-h MV (n = 6). *P ≤ 0.05 vs. SB.

DISCUSSION

This is the first investigation to demonstrate a substantial blunting of both endothelium-dependent and -independent vasodilation in resistance vessels of the diaphragm following prolonged MV. These rapid decrements in vasomotor control provide a mechanism for the altered diaphragm hyperemic response following prolonged MV (10).

Specifically, following prolonged MV there was a diminished endothelium-dependent vasodilation in response to flow and the muscarinic agonist ACh. Interestingly, in the presence of l-NAME, differences in ACh-induced vasodilation between groups were abolished, suggesting that the loss of NO bioavailability with prolonged MV contributes to the diminished dilation. This notion is supported by the significant reduction in eNOS mRNA of the diaphragm arterioles after prolonged MV. However, contrary to our hypothesis, there was a diminished vasodilation to SNP, demonstrating that the blunted dilation of diaphragm resistance vessels with prolonged MV extends beyond the endothelium and induces alterations within the smooth muscle. Altered mechanical/material properties of the vessels with prolonged MV could contribute to these findings; however, no gross differences were observed in the passive pressure-diameter responses. Contrary to our hypothesis regarding enhanced contractile responses, there were no differences in the myogenic responses between groups. Furthermore, prolonged MV did not alter vasoconstrictor responses to NE; however, there was a diminished vasoconstriction to PE. In addition, vessels from the MV group displayed reduced sensitivity to both α-adrenergic agonists.

Reduced Endothelium-Dependent and -Independent Vasodilation

After prolonged MV, the blunted dilation of arterioles is due to both endothelium-dependent and -independent mechanisms (Fig. 3, A and B). The reduced endothelium-dependent vasodilation with prolonged MV may result from one or a combination of mechanisms, including 1) reduced NO bioavailability, 2) increased levels of inflammatory cytokines, and/or 3) oxidative stress. Vascular smooth muscle relaxation is highly dependent on the activity of eNOS and the release of bioactive NO. Importantly, shear stress is a key modulator of short- and long-term endothelial structure and function, including eNOS activity (46, 47). In our study, following prolonged MV, flow-induced dilation was reduced by 40% (Fig. 1) with a similar blunting of ACh-mediated vasodilation (Fig. 2). This is an intriguing finding, considering the diaphragm possesses a substantial vasodilatory reserve, even at maximal exercise (33). Woodman and colleagues (46), using isolated soleus feed arteries from senescent rats, demonstrated that increased shear stress can upregulate eNOS mRNA expression in as little as 4 h. Therefore, the reduced dilation in response to increasing intraluminal flow (Fig. 1) may be a consequence of the rapid reduction in blood flow with MV (10) and presumably shear stress, acting on the endothelium of the quiescent diaphragm resistance vessels and possibly reducing eNOS activity. Indeed, we observed a significant reduction in eNOS mRNA expression of diaphragm arterioles with MV (Fig. 2C). Although eNOS mRNA expression is not a measure of eNOS activity per se, it does suggest the reduced diaphragmatic blood flow as a possible signaling mechanism for altering eNOS expression and, consequently, NO bioavailability.

Inflammatory cytokines, such as tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6) have been shown to alter eNOS mRNA expression (3) posttranscriptionally and/or posttranslationally via the NF-κB pathway (31). Furthermore, IL-6 may act downstream to reduce NO bioavailability and increase superoxide anion (O₂^·−) production (16, 25). Interestingly, in a model of hindlimb disuse atrophy, TNF-α and IL-6 mRNA are upregulated (12). In addition, inflammatory cytokines have the ability to downregulate eNOS and inhibit NO release (3). Since the diaphragm is inactive during controlled MV, IL-6 may be involved in the endothelial dysfunction seen in diaphragm arterioles, similar to disuse atrophy. Elevated levels of inflammatory cytokines may play a key role with prolonged MV by reducing NO bioavailability through downregulation of eNOS and increased ROS production. With respect to controlled MV, we (7) and others (18, 40) have previously demonstrated elevations in both proinflammatory cytokines and oxidative stress such as IL-6 and TNF-α. Recently, our group (10) demonstrated a severely reduced diaphragm microvascular oxygen partial pressure (Po_2m) at rest and during contractions following prolonged MV. Furthermore, reduced oxygenation has been shown to increase electron availability and enhance O₂^·− production (9). The reduced diaphragm Po_2m (10), and possible elevations of O₂^·−, could explain in part the blunted endothelium-dependent and -independent dilation with prolonged MV. Interestingly, it has been suggested that uncoupled eNOS is a primary source of intracellular ROS generation in endothelial cells (23). In addition, superoxide dismutase (SOD), an enzyme that catalyzes the dismutation of O₂^·− into hydrogen peroxide (H₂O₂), can readily scavenge bioactive NO in pathological conditions (47). During prolonged MV, it is likely that eNOS is uncoupled in the presence of increased levels of O₂^·−, working in coordination with SOD to degrade NO, and thus attenuating endothelium-dependent dilation.

The blunted vasodilation with prolonged MV in diaphragm arterioles is not due solely to endothelial dysfunction. Although we observed no gross structural alterations in these arterioles (i.e., no difference in the passive pressure response), there may be alterations in vascular mechanical properties following MV (i.e., stress-stretch relationship). Functional alterations, such as reduced activity of cGMP and/or intracellular Ca²⁺ release, may be present in vascular smooth muscle cells. Furthermore, the formation of ROS has been shown to reduce the action of the NO-cGMP pathway in vascular smooth muscle relaxation (47), which could explain the blunted endothelium-independent relaxation with SNP (Fig. 2D). Therefore, increased oxidative stress (O₂^·− and ROS) and proinflammatory cytokines with prolonged MV could impact multiple pathways of NO production (e.g., downregulation of eNOS) and availability via scavenging by O₂^·− and potentially H₂O₂, all of which would culminate in a diminished vasodilatory capacity as demonstrated herein.

Altered α-Adrenergic Responses

α-Adrenergic receptors (α₁- and α₂-receptors) are expressed in resistance vasculature and mediate vasoconstriction by inducing vascular smooth muscle contraction. We examined α₁- and α₂-adrenergic receptor contributions to vasoconstriction in diaphragm 1A via NE and PE, a nonselective α-adrenergic agonist and α₁-specific agonist, respectively. After prolonged MV, diaphragm arterioles showed no difference in vasoconstriction to NE compared with the SB group; yet, contrary to our hypothesis, they displayed a 20% reduction in vasoconstriction in response to PE (Fig. 5B), suggesting a possible reduction in α₁-mediated vascular control. Albeit the exact mechanism for the reduced α₁-mediated vasoconstriction produced in this study was not determined, prior evidence suggests that the location and sensitivity of the adrenergic receptors modulate vascular tone (1, 22). Aaker and Laughlin (1) found that second-order arterioles (2A) in the diaphragm are primarily regulated via α₁-receptors, with little contribution from α₂-receptors. The diaphragm has an enormous capacity for flow (33), similar to other highly oxidative skeletal muscle (4) but dissimilar in that it is never truly quiescent in health. Furthermore, the diaphragm demonstrates a specialized microcirculation [e.g., countercurrent capillary flow (20)], structurally and functionally optimized to facilitate oxygen transfer (34). However, the diaphragm is particularly susceptible to ischemia-reperfusion damage and associated downward shifts in the force-frequency relationship (42). Therefore, it could be hypothesized that the diaphragm may demonstrate a blunted vasoconstriction in health to neural/humoral stimuli vs. other skeletal muscle to maintain O₂ delivery. The literature is relatively scant on this topic. However, Supinski et al. (43) demonstrated NE infusion increases blood flow to the diaphragm in dogs. Furthermore, contrary to the robust constriction to these agonists in arterioles of skeletal muscle (21, 22, 38), diaphragm arterioles are less responsive to α-adrenergic constriction (1). The blunted α₁-adrenergic and unaltered α₂-adrenergic vasoconstriction following MV may stem from downstream G protein coupling, second-messenger signaling (i.e., inositol 1,4,5-trisphosphate), and/or intracellular Ca²⁺ release. Since α₁- and α₂-mediated smooth muscle contraction works via different downstream G protein coupling and second messenger signaling (8, 48), this could explain, in part, the unaltered α₂-response with a reduced sensitivity to both agonists.

Ramifications of Rapid Vasomotor Dysfunction in the Diaphragm with MV

To sustain diaphragmatic metabolic and contractile function, there must be a tight coupling of O₂ delivery to myocyte energetics and O₂ consumption (32, 34). During weaning from MV, there is fivefold increase in diaphragm venous blood lactate, a ~30% decrease in diaphragm glycogen stores, and 50% reduction in ATP and CP (24), all of which indicate an inability to match O₂ delivery to O₂ consumption, similar to the diminished microvascular oxygenation as previously observed (10). Therefore, blunted dilation of diaphragm arterioles after MV appears to be playing an integral role in the inability to augment blood flow to the diaphragm during contractions and predicates weaning failure.

Limitations

The SB group was not subject to the same duration of anesthesia; however, if this was an anesthesia issue it would have to be selective to the diaphragm, as other skeletal muscle does not demonstrate reductions in blood flow over the 6-h period. No data that we are aware of, demonstrate the diaphragm to be the selective target of systemic anesthesia. Our investigations were in the absence of in vivo neural and humoral influences, which is common for all isolated vessel preparations. In the current study, we used whole-log doses of ACh and SNP from 10⁻⁹ to-10⁻⁴M, which typically results in a plateau in relaxation at the higher doses (13, 41). However, as apparent in Fig. 2, a clear plateau was not evident. Nonetheless, the maximal response elicited by these agonists at these doses was clearly blunted after MV (Fig. 2). Diaphragm quiescence was not verified directly; however, the diaphragm is inactive in anesthetized rats during prolonged controlled MV and has been verified through electromyography (35, 37). As Pco₂was maintained ≤30 mmHg, there was likely no sensory input to the respiratory control center by carotid chemoreceptor activation. In addition, pleural pressures were not determined during MV and could have affected diaphragmatic blood flow. However, previous work demonstrates that hyperinflation of the lungs fails to significantly alter diaphragmatic blood flow (19). Finally, we did not quantify α-receptor mRNA expression in either group, but it is possible that α-receptor mRNA is reduced similarly to the reduced eNOS mRNA expression presented herein, but this requires future investigation.

Conclusion

Following 6 h of MV, vascular alterations occur through both endothelium-dependent and -independent pathways. This is the first study, to our knowledge, demonstrating that diaphragm arteriolar dysfunction is associated with prolonged MV and likely contributes to VIDD and weaning difficulties. Further studies are needed to investigate the possible role of other vasoactive mediators such as angiotensin II and endothelin-1, which may be upregulated with MV, as well as the potential contribution of oxidative stress to vasomotor alterations with prolonged MV.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant1 R15 HL137156-01A1. C. S. Bruells was supported by the Deutsche Forschungsgemeinschaft (BR 3998/1-1).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

A.G.H., D.C.P., and B.J.B. conceived and designed research; A.G.H., R.T.D., D.R.B., O.N.K., C.S.B., D.J.M., A.B.O.-A., and B.J.B. performed experiments; A.G.H., R.T.D., and B.J.B. analyzed data; A.G.H., R.T.D., D.R.B., C.S.B., D.J.M., A.B.O.-A., D.C.P., and B.J.B. interpreted results of experiments; A.G.H. and R.T.D. prepared figures; A.G.H. drafted manuscript; A.G.H., R.T.D., D.R.B., O.N.K., C.S.B., D.J.M., A.B.O.-A., D.C.P., and B.J.B. edited and revised manuscript; A.G.H., R.T.D., D.R.B., O.N.K., C.S.B., D.J.M., A.B.O.-A., D.C.P., and B.J.B. approved final version of manuscript.