**1. Introduction**

Microvascular endothelial barrier disruption occurs in a large number of disease states, such as stroke, sepsis, diabetes, hereditary and acquired angioedema, commonly induced by a variety of endogenous inflammatory mediators such as bradykinin [1–6]. Novel therapeutic approaches to prevent or reduce microvascular permeability are paramount to avoid tissue edema and to maintain sufficient blood supply to target organs. In this context, statins have been described to reduce vascular permeability and edema formation in different animal and clinical studies [7–10], yet the underlying mechanisms have not been investigated in an intact muscle microvasculature.

Bradykinin has several pathophysiological functions and activates the B2 receptor, which is constitutively expressed on the vasculature and increases vascular permeability in post-capillary venules [11]. Moreover, bradykinin is an important mediator in stroke, sepsis, diabetes, hereditary and acquired angioedema [1–6]. Bradykinin may also play a key role in the vascular leakage and pulmonary edema in patients with COVID-19 [12–14]. Angiotensin converting enzyme 2 (ACE2) has been implicated as the cellular receptor of SARS-CoV-2 virus [15,16], and reduced ACE2 activity may indirectly activate the kallikrein–bradykinin pathway to increase vascular permeability [17].

In vitro and in vivo studies have shown that the increase in vascular permeability induced by bradykinin depends on the generation of reactive oxygen species [18,19]. We previously reported that bradykinin-induced microvascular permeability in the brain pial microvasculature in vivo is directly associated with the release of reactive oxygen species following bradykinin receptor activation [20]. The pro-inflammatory cytokine IL-1β has been shown potentiate the actions of bradykinin and to increase microvascular permeability and edema formation after experimental cerebral ischemia reperfusion injury [19,21,22]. Under ischemic conditions, IL-1β is rapidly released from brain tissue, leading to NADPH oxidase assembly and activation, which then rapidly potentiates the permeability response to bradykinin [19]. Notably, potentiation of bradykinin-induced increases in cerebral microvascular permeability are blocked by the IL-1 receptor antagonist, IL-1ra [19]. Moreover, acute release of IL-1β has been described as a key inflammatory event in patients with COVID-19 [23–25] that could also potentiate bradykinin-induced vascular permeability.

Clinical and experimental studies indicate several beneficial effects of statins independent of their cholesterol-lowering action [26–28]. Statins may have the potential to reduce oxidative stress by modulating Nrf2-regulated antioxidant genes [29,30], such as heme oxygenase 1 (HO-1) known to afford protection in rodent models of ischemia in vivo [31,32] and in vascular cells in vitro [29,33]. Further evidence suggests that simvastatin may upregulate HO-1 independently of Nrf2 [34].

To date there are no studies focused on the protective actions of statins against IL-1β mediated potentiation of bradykinin-induced microvascular permeability. In this study, we investigate for the first time the effects of pretreatment of rats with simvastatin on bradykinin- and IL-1β-induced microvascular permeability using intravital microscopy in an intact cremaster muscle preparation that to date has not been reported. Our findings suggest that simvastatin prevents microvascular hyperpermeability induced by IL-1β and bradykinin via inhibition of NADPH oxidase and inhibition of reactive oxygen species generation.

#### **2. Materials and Methods**

#### *2.1. Animals and Isolation of the Cremaster Skeletal Muscle Preparation*

This study conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996) and is in accordance with UK Home Office regulations (Animals Scientific Procedures) Act, 1986. Approved by UK Home Office Animal Project License (PPL Number: 70/8934).

Male Wistar rats (Charles River, UK), 4–6 weeks old and weighing 80–100 g), were killed by exposure to a rising concentration of CO<sup>2</sup> followed by cervical dislocation. A longitudinal midline incision (1–2 cm) was made along the abdomen to expose the underlying organs. All the small branches of the aorta except the common iliac arteries leading branches that did not supply the chosen cremaster were tied off and the vena cava was then punctured to create an outlet for the blood that was flushed out of the circulation. The aorta was cannulated orthogradely with a polythene tubing (outside diameter 0.61 mm). The left common iliac and the right femoral and internal iliac arteries were ligated to ensure that perfusion was directed to the right external iliac artery supplying the cremaster artery and the cremaster muscle microvasculature. The tissue was perfused with a modified St. Thomas' cardioplegic

solution (mM: 10 MgCl2, 110 NaCl, 8 KCl, 1 CaCl2, 10 HEPES) [35] containing heparin (30 U/mL) and isoproterenol 10 µM buffered to pH 7.0 ± 0.05 for 10 min.

#### *2.2. Superfusion of Cremaster Muscle Preparation*

A longitudinal incision was made along the length of the ventral aspect of the scrotum and the overlying fascia and connective tissue were carefully removed. The cremaster was pulled out with the testicle using a pair of blunt forceps, and the distal end of the muscle was secured on a SylGard block using histology pins (Watkins and Doncaster, Kent, England). The intact cremaster preparation was then transferred to a modified stage of an intravital microscope (ACM, Zeiss, Oberkochen, Germany) and continuously superfused (2 mL·min−<sup>1</sup> ) with an albumin-free Krebs solution (pH 7.4) gassed with 5% CO<sup>2</sup> in air and maintained at 37 ◦C. The superfusate contained the Na<sup>+</sup> channel blocker lidocaine (20 mg·L −1 ) to block neural activity and to minimize cremaster muscle contractions.

#### *2.3. Measurement of Post-Capillary Venule Permeability to FITC-Albumin*

The stabilizing solution perfusing the cremaster vasculature was replaced with Krebs solution (mM: 118 NaCl; 4.7 KCl; 2.52 CaCl2; 1.18 MgSO4.7H2O; 1.18 KH2PO4; 25 NaHCO3, 5 glucose and buffered to pH 7.4 <sup>±</sup> 0.05) containing bovine albumin (10 mg·mL−<sup>1</sup> ) delivered by a gravity controlled reservoir at 0.5 mL·min−<sup>1</sup> . After 30 min, Krebs perfusion of the vasculature was stopped and a bolus of Krebs solution containing FITC-albumin (5 mg·mL−<sup>1</sup> ) was injected into the perfusion line. Post-capillary venules were identified by noting the direction of flow, as the microvasculature was filled with the fluorescent dye, using a 10× water immersion objective (numerical aperture 0.5). Images were captured using an FITC filter cube (Chroma Technology Bellows Falls, VT, USA) via image-intensified CCD camera (Photonic Sciences, Robertsbridge, E. Sussex, UK) for subsequent analysis (ImageHopper; Samsara Research, Dorking, Surrey, UK).

Perfusion pressure was lowered to atmospheric, and pressure differences in the vasculature were allowed to dissipate over the course of 3 min. In previous studies, we have demonstrated a linear correlation between the light collected with the dye concentration and as well as with the square of the diameter of the microvessel up to a 60 µm limit [36]. Permeability measurements were obtained from an image sequence acquired at 1 s intervals over 100 s. The dye concentration difference across a vessel was calculated from the difference between the regions of interest positioned on an image stack (see Figure 1A,B). Permeability was determined from the rate of decrease in that difference, obtained by fitting an exponential to the data (Figure 1C) such that P = kD/4, where k is the rate constant and r is the vessel diameter. The lack of axial flow under the experimental conditions was confirmed by viewing fluorescent microspheres (1 µm diameter) within the vasculature (data not shown). It was possible to generate a permeability map on a few occasions when the venule was on the exposed surface of the cremaster preparation, so that there was no overlying tissue and that any escaped dye dissipated rapidly. Under these circumstances, the rate constants could be calculated on a pixel by pixel basis during the exponential fall of dye (see Figure 1D) by taking linear regression of the log (V-I), where V and I are the pixel values within the vessel and the interstitium, respectively.

#### *2.4. Role of Nitric Oxide and Reactive Oxygen Species in Microvascular Permeability*

To determine the role of nitric oxide and reactive oxygen species on basal permeability, the cremaster preparation was superfused for 5 min with a nitric oxide synthase (NOS) inhibitor, N-ω-nitro-L-arginine methyl ester (L-NAME; 10 µM) and/or the free radical scavengers superoxide dismutase (SOD, 100 U·mL−<sup>1</sup> ) and catalase (CAT, 100 U·mL−<sup>1</sup> ). Further experiments examined the effects of the vasoactive mediators histamine (1 µM) and bradykinin (100 nM) on permeability in the absence or presence of L-NAME or SOD and CAT.

*Antioxidants* **2020**, *9*, x FOR PEER REVIEW 4 of 17

**Figure 1.** Basal and agonist stimulated permeability measurements in single venules in a rat cremaster muscle preparation. (**A**) Representative fluorescence image of the cremaster microvasculature following arterial FITC-albumin infusion. A sequence of images was captured at 1 s intervals after all axial flow had ceased, during which time (70–110 s, panel **B**) either histamine or bradykinin was applied topically. (**B**) Image stack was analyzed by placing regions of interest (ROIs) over the 33 µm diameter venule (red) and the neighboring interstitium (green). (**C**) Difference between these values for the two ROI yields the albumin concentration gradient across the microvessel. The rate constant (k) for the fitted monoexponential and the diameter gives the permeability value P = kD/4, assuming a circular diameter. (**D**) A few venules, such as the one illustrated in panel A, were on the surface of the cremaster, not overlaid with skeletal muscle fibers, which allowed a color-coded permeability map to be generated: the scale values are expressed as cm·s<sup>−</sup>1 × 10−6. The left-hand image was generated following application of 10 nM bradykinin and the right-hand image after 100 nM bradykinin. **Figure 1.** Basal and agonist stimulated permeability measurements in single venules in a rat cremaster muscle preparation. (**A**) Representative fluorescence image of the cremaster microvasculature following arterial FITC-albumin infusion. A sequence of images was captured at 1 s intervals after all axial flow had ceased, during which time (70–110 s, panel **B**) either histamine or bradykinin was applied topically. (**B**) Image stack was analyzed by placing regions of interest (ROIs) over the 33 µm diameter venule (red) and the neighboring interstitium (green). (**C**) Difference between these values for the two ROI yields the albumin concentration gradient across the microvessel. The rate constant (k) for the fitted monoexponential and the diameter gives the permeability value P = kD/4, assuming a circular diameter. (**D**) A few venules, such as the one illustrated in panel A, were on the surface of the cremaster, not overlaid with skeletal muscle fibers, which allowed a color-coded permeability map to be generated: the scale values are expressed as cm·s <sup>−</sup><sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>6</sup> . The left-hand image was generated following application of 10 nM bradykinin and the right-hand image after 100 nM bradykinin.

#### To determine the role of nitric oxide and reactive oxygen species on basal permeability, the *2.5. Bradykinin- and IL-1*β*-Induced Increases in Microvascular Permeability*

*2.4. Role of Nitric Oxide and Reactive Oxygen Species in Microvascular Permeability* 

cremaster preparation was superfused for 5 min with a nitric oxide synthase (NOS) inhibitor, N-ωnitro-L-arginine methyl ester (L-NAME; 10 µM) and/or the free radical scavengers superoxide dismutase (SOD, 100 U·mL−1) and catalase (CAT, 100 U·mL−1). Further experiments examined the effects of the vasoactive mediators histamine (1 µM) and bradykinin (100 nM) on permeability in the absence or presence of L-NAME or SOD and CAT. *2.5. Bradykinin- and IL-1β-Induced Increases in Microvascular Permeability*  Increasing concentrations of bradykinin (10−9, 10−8 and 10−7 M) were applied abluminally to the Increasing concentrations of bradykinin (10−<sup>9</sup> , 10−<sup>8</sup> and 10−<sup>7</sup> M) were applied abluminally to the cremaster muscle to elicit dose dependent increases in FITC-albumin permeability. After a dose–response curve to bradykinin, the cremaster muscle was rapidly superfused and IL-1β (30 pM) applied abluminally for 10 min. The preparation was then superfused to remove IL-1β, and a new dose–response curve to bradykinin (10−<sup>9</sup> , 10−<sup>8</sup> and 10−<sup>7</sup> M) performed. To reduce variability between drug applications, the same region of the cremaster microvasculature was observed throughout an entire experiment, paired permeability measurements obtained in single post-capillary venules.

#### cremaster muscle to elicit dose dependent increases in FITC-albumin permeability. After a dose– response curve to bradykinin, the cremaster muscle was rapidly superfused and IL-1β (30 pM) *2.6. Inhibition of NADPH Oxidase Assembly*

applied abluminally for 10 min. The preparation was then superfused to remove IL-1β, and a new dose–response curve to bradykinin (10−9, 10−8 and 10−7 M) performed. To reduce variability between To determine the involvement of NADPH oxidase in the microvascular hyperpermeability induced by IL-1β and bradykinin, the cremaster preparation was superfused for 10 min with IL-1β (30 pM) in the presence of apocynin (Apo, 1 µM), a specific inhibitor of NADPH oxidase in control rats as well as in simvastatin pretreated rats. The preparation was then rapidly superfused to remove IL-1β and apocynin and bradykinin (100 nM) applied abluminally.

#### *2.7. Pretreatment of Animals with Simvastatin*

Simvastatin (5 mg·kg−<sup>1</sup> ) was administered to rats intraperitoneally 24 h before isolation of the cremaster muscle preparation.

#### *2.8. Inhibition of Heme Oxygenase-1 with Tin Protophoryrin IX (SnPP)*

The HO-1 inhibitor, tin protoporphyrin IX (SnPP) (5 µM), was applied abluminally for 10 min. The preparation was then superfused to remove the (SnPP), and bradykinin (100 nM) was applied. The cremaster muscle was then rapidly superfused (washed) and IL-1β (30 pM) was co-applied with (SnPP) (5 µM) abluminally for 10 min. The preparation was then superfused to remove IL-1β and (SnPP) and bradykinin (100 nM) applied abluminally.

#### *2.9. Reagents*

All chemicals were purchased from Sigma-Aldrich (Dorset, UK).

#### *2.10. Statistical Analysis*

*Pathways* 

Experimental data represent paired permeability measurements in single venules from different animals or are expressed as mean ± SEM of measurements in single venules from *n* = 4–10 animals. Data were analyzed using a paired or unpaired Student's *t*-test and ANOVA in GraphPad Prism 6.0 (La Jolla, CA, USA), with *p* <0.05 considered statistically significant.

#### **3. Results**

#### *3.1. Role of NO and Reactive Oxygen Species in Modulating Basal Microvascular Permeability*

Application of the nitric oxide synthase inhibitor L-NAME (10 µM) increased permeability (0.69 ± 0.26 cm·s <sup>−</sup><sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>6</sup> , *p* < 0.05) above basal levels (0.33 ± 0.23 cm·s <sup>−</sup><sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>6</sup> , Figure 2). Notably, co-application of superoxide dismutase (SOD, 100 U·mL−<sup>1</sup> ) and catalase (CAT, 100 U·mL−<sup>1</sup> ) with L-NAME abrogated the permeability increase (0.15±0.09 cm·s <sup>−</sup><sup>1</sup> <sup>×</sup>10−<sup>6</sup> vs. 0.25±0.05 cm·<sup>s</sup> <sup>−</sup><sup>1</sup> <sup>×</sup>10−<sup>6</sup> , Figure 2) evoked by L-NAME. *Antioxidants* **2020**, *9*, x FOR PEER REVIEW 6 of 17

**Figure 2.** Constitutive nitric oxide reduces basal post-capillary venule permeability. Inhibiting constitutive eNOS with L-NAME (10 µM) resulted in a significant permeability increase, while scavenging reactive oxygen species with a combination superoxide dismutase and catalase (100 U·mL<sup>−</sup>1 each) reduced basal permeability. When superoxide dismutase and catalase were co-applied with L-NAME, there was no permeability change. Data from paired measurements in 4 venules from 4 different animals. Data were analyzed using a paired Student's *t*-test. **Figure 2.** Constitutive nitric oxide reduces basal post-capillary venule permeability. Inhibiting constitutive eNOS with L-NAME (10 µM) resulted in a significant permeability increase, while scavenging reactive oxygen species with a combination superoxide dismutase and catalase (100 U·mL−<sup>1</sup> each) reduced basal permeability. When superoxide dismutase and catalase were co-applied with L-NAME, there was no permeability change. Data from paired measurements in 4 venules from 4 different animals. Data were analyzed using a paired Student's *t*-test.

and catalase (Figure 3A). In contrast, as shown in Figure 3B, the permeability increase induced by bradykinin (100 nM, 2.2 ± 0.2 cm·s−1 × 10−6) was unaffected by L-NAME (1.6 ± 0.4 cm·s−1 × 10−6) but blocked by co-application of superoxide dismutase and catalase (100 U·mL−1; −0.1 ± 0.1 cm·s−1 × 10−6).

*3.2. Histamine- and Bradykinin-Induced Microvascular Permeability Is Mediated by Different Signaling* 
