Acute lysyl oxidase inhibition alters microvascular function in normotensive but not hypertensive men and women

Abstract

The lysyl oxidase (LOX) family of enzymes regulates collagen cross-linking. LOX is upregulated in hypertension, increasing vascular stiffness. In vivo human research is sparse, as long-term LOX inhibition in animals causes vascular instability. Our aim was to evaluate the effects of LOX inhibition on cutaneous microvascular function to determine whether LOX function was upregulated in hypertensive humans. Four intradermal microdialysis fibers were placed in the forearm of 10 young [age: 24 ± 1 yr, mean arterial pressure (MAP): 87 ± 2 mmHg], 10 normotensive (age: 50 ± 2 yr, MAP: 84 ± 1 mmHg), and 10 hypertensive (age: 53 ± 2 yr, MAP: 112 ± 2 mmHg) subjects. Two sites were perfused with 10 mM β-aminopropionitrile (BAPN) to inhibit LOX. The remaining two sites were perfused with lactated Ringer solution (control). A norepinephrine dose response (10⁻¹²−10⁻² M) was performed to examine receptor-mediated vasoconstrictor function. A sodium nitroprusside dose response (10⁻⁸−10^−1.3 M) was performed to examine vascular smooth muscle vasodilator function. Red blood cell flux was measured via laser-Doppler flowmetry and normalized to cutaneous vascular conductance (flux/MAP). LogEC₅₀ values were calculated to determine changes in vasosensitivity. Skin tissue samples were analyzed for both extracellular matrix-bound and soluble LOX. LOX inhibition augmented vasoconstrictor sensitivity in young (control: −6.0 and BAPN: −7.1, P = 0.03) and normotensive (control: −4.8 and BAPN: −7.0, P = 0.01) but not hypertensive (control: −6.0 and BAPN: −6.1, P = 0.79) men and women. Relative to young subjects, extracellular matrix-bound LOX expression was higher in hypertensive subjects (young: 100 ± 8 and hypertensive: 162 ± 8, P = 0.002). These results suggest that upregulated LOX may contribute to the vascular stiffness and microvascular dysfunction characteristic in hypertension.

NEW & NOTEWORTHY Matrix-bound lysyl oxidase (LOX) and LOX-like 2 expression are upregulated in the microvasculature of hypertensive men and women. Microvascular responsiveness to exogenous stimuli is altered with localized LOX inhibition in healthy men and women but not hypertensive adults. The LOX family differentially affects microvascular function in hypertensive and normotensive men and women.

INTRODUCTION

Hypertension is a highly prevalent chronic disease affecting approximately one-third of adults (18), which increases the risk of developing coronary artery disease, renal failure, and stroke (44). Among the many hypertension-associated pathological alterations, increased vascular stiffness is a hallmark maladaptation contributing to end-organ damage (33, 40, 59). Many mechanisms contribute to increased vascular stiffness, including angiotensin II-mediated vascular smooth muscle hypertrophy and hyperplasia, advanced glycation end products, increased oxidant stress, proinflammatory cytokines, and the matrix remodeling enzyme lysyl oxidase (LOX) (28, 42, 49, 62).

LOX is a copper-dependent amine oxidase that regulates vascular structure by catalyzing collagen and elastin cross-linking. LOX oxidatively deaminates lysine and hydroxylysine residues to reactive semialdehydes. The reactive semialdehydes then spontaneously condense to form covalent cross-links, increasing vessel stiffness (42). Four similar enzymes to LOX, LOX-like (LOXL)1–4 enzymes, have been identified with varying levels of expression in different tissues. LOX expression is requisite for the proper development and maintenance of vascular structure. LOX knockout mice fail to develop fully formed vascular and respiratory systems and are not viable (25, 32), and prolonged systemic inhibition of LOX decreases vessel integrity, leading to aneurysm development (29, 37, 52). In humans, depressed expression of LOX leads to spontaneous coronary artery dissection (51). Because of the dangerous outcomes associated with systemic LOX inhibition, there are currently no human studies evaluating the direct effects of inhibition of the entire LOX family on vascular function or structure. There is currently only one phase I study that has recently been completed with a systemic LOXL2-specific inhibitor (46).

Despite the integral role of LOX in the maintenance of proper vessel structure, overexpression of LOX is associated with cardiovascular pathologies and contributes to cardiac and vascular stiffening (31), indicating that LOX may be an emerging therapeutic target. In murine models of hypertension, LOX activity, LOX expression, and the resulting collagen cross-linking are increased (13, 19). In contrast, collagen cross-linking and fibrosis are attenuated with systemic LOX inhibition (38, 45). In limited human studies, LOX expression and collagen cross-linking are elevated in both men and women with chronic heart failure (31) and pulmonary hypertension (38). However, to date, the mechanistic roles of LOX in the human microcirculation are unknown.

The human cutaneous circulation is an accessible model to assess mechanisms underlying systemic microvascular dysfunction (24). Hypertension-associated pathological alterations in vascular structure and endothelial dysfunction are observable in the cutaneous microvasculature (1, 12, 15, 55). Moreover, both LOX and LOXL2 are expressed in the skin microcirculation (17, 31, 35, 39). Skin-specific approaches (intradermal microdialysis) (10) allow for the examination of the effects of LOX inhibition in the human microvasculature without the risk of compromising the structural integrity of the large arteries, allowing for the first time direct LOX inhibition in human microvasculature.

We have previously observed attenuated cutaneous vasodilation in response to high concentrations of the endothelium-independent vasodilator sodium nitroprusside (SNP) in hypertensive men and women (12); however, the mechanisms mediating this attenuated vasodilator response have not been fully elucidated. It is possible that structural and/or functional changes induced by LOX overexpression in hypertension play a role. Hypertension is characterized by a proconstrictor state. LOX inhibition has been shown to attenuate norepinephrine (NE)-induced vasoconstriction in isolated carotid arteries from rats (9). However, whether LOX inhibition attenuates NE-induced vasoconstriction in hypertensive men and women has not been established.

To evaluate the effects of LOX on microvascular function, we recruited essential hypertensive men and women and age-matched normotensive men and women to established the relation between LOX expression, microvascular function, and blood pressure status. We also recruited a group of young normotensive men and women to control for the appearance of minor cutaneous microvascular dysfunction that is apparent in healthy middle-aged adults (6).

The aim of this study was to assess the effects of acute LOX and LOXL1–4 inhibition on human cutaneous microvascular function as a first step in understanding the vascular roles that the LOX family of enzymes play in the microvasculature in vivo. Specifically, the purpose of this study was to measure soluble and extracellular matrix-bound LOX/LOXL2 expression in skin tissue samples and to determine changes in vascular smooth muscle vasodilator and vasoconstrictor reactivity in response to acute localized LOX inhibition in young, middle-aged normotensive, and essential hypertensive men and women. We hypothesized that 1) expression of extracellular matrix-bound LOX and LOXL2 would be upregulated in skin samples from hypertensive human subjects and 2) inhibition of the LOX family of enzymes would augment vascular smooth muscle-mediated vasodilation and attenuate vasoconstriction in hypertensive men and women.

METHODS

Ethical approval.

Experimental protocols were approved by the Institutional Review Board of The Pennsylvania State University and conformed with the Declaration of Helsinki. Written and verbal consent were obtained voluntarily from all participants before participation in the study.

Participants.

Participants completed a medical screening including a review of their health history, blood pressure screening, and blood lipid and chemistry analysis (Quest Diagnostics). Middle-aged (normotensive and hypertensive) participants also underwent blood pressure assessment with a 24-h ambulatory monitor to confirm blood pressure status.

Three groups of participants were recruited to participate: young, essential hypertensive, and age-matched normotensive. To qualify, young subjects must have been between 18 and 30 yr old, whereas essential hypertensive and age-matched normotensive subjects must have been between 45 and 70 yr old. Hypertensive status was confirmed according to JNC7 guidelines, with each participant presenting with a seated systolic blood pressure (SBP) ≥ 140 mmHg or diastolic blood pressure (DBP) ≥ 90 mmHg on at least two separate occasions (8). Blood pressure status was confirmed through use of a 24-h ambulatory blood pressure monitor (Ambulo 2400, Mortara Instruments) that measured blood pressure every 30 min during the waking hours (0800–2200 hours) and every 60 min during sleep (2200–0800 hours). Due to the inclusion of sleeping hours, mean 24-h blood pressure measured from ambulatory monitors is usually less than what is observed during seated blood pressure measurement. Thus, hypertensive status was confirmed from ambulatory monitor data if 24-h average SBP was ≥130 mmHg and/or DBP was ≥80 mmHg, as directed by European Society of Hypertension/European Society of Cardiology guidelines (34).

Other than high blood pressure, participants were included if they were apparently healthy, did not smoke, and were not taking medications with vascular effects. All premenopausal women were normally menstruating and were studied during either the early follicular phase (days 1–7) or the placebo phase if they were taking hormonal birth control. Postmenopausal women were not taking hormone replacement therapy. All participants refrained from consuming alcohol and caffeine and from exercise for 12 h before the study.

In vitro analysis.

Forearm skin tissue samples were obtained via punch biopsy from a subset of subjects (8 young, 9 normotensive, and 9 hypertensive), as previously described (55). Skin samples were homogenized in 200 μl RIPA containing protease inhibitors and centrifuged at 12,000 rpm to recover supernatant. Pellet was solubilized in SDS-containing buffer (1× Laemmli buffer). Western blot analysis was performed to investigate DOC-soluble (soluble, non-matrix-bound) and SDS-soluble (extracellular matrix-bound) LOX. Anti-LOX (PA1-46020, Invitrogen) (56) and anti-LOXL2 (ab179810, Abcam) (60) antibodies were used to measure expression levels of LOX and LOXL2. Anti-GAPDH (NB300-221, Novus) was used to obtain control for densitometry normalization of soluble LOX/LOXL2.

In vivo analysis.

Intradermal microdialysis is a minimally invasive technique that allows for the delivery of small-molecular-weight substances to a discrete area of the cutaneous microvasculature without eliciting systemic effects (24). We have used this technique previously in similar subjects without complication (11, 53, 55). Four intradermal microdialysis fibers (55-kDa cutoff, CMA Microdialysis) were placed into the dermal layer of the ventral forearm, as previously described (55, 57). Pharmacological agents were mixed just before use, dissolved in lactated Ringer solution, and filtered with a syringe microfilter (0.2-μm membrane, Acrodisc). Microdialysis sites were randomized into two pairs. One microdialysis site in each pair received a 2-h pretreatment with 10 mM β-aminopropionitrile (BAPN; Sigma-Aldrich) to nonspecifically inhibit LOX and all LOXL enzymes, and the other site was perfused with lactated Ringer solution to serve as the control. Pilot testing on isolated tissue found that the IC₅₀ for BAPN was 2.5 mM; 10 mM BAPN was chosen to overcome any local delivery issues associated with intradermal microdialysis. At the time of this experiment, the LOXL2-specific inhibitor (PAT-1251) was not available for use in humans. In vivo pilot testing for BAPN confirmed that a 10 mM concentration administered via microdialysis was efficacious and did not induce a baseline shift in vasodilation and that 2 h was sufficient for BAPN to take effect. Pharmacological agents were perfused through the fibers at a rate of 2 μl/min (BASi Bee Hive controller and Baby Bee syringe drive). The 2-h pretreatment with BAPN provided sufficient time from the resolution of hyperemia from microdialysis fiber placement. Local heating units (Moor Instruments) were placed over each microdialysis site and were set to thermoneutral skin temperature (33°C). A laser-Doppler flowmeter probe (Moor Instruments) was placed in each local heating unit to continuously measure red blood cell flux, a relative measure of skin blood flow. Each pair of microdialysis fibers was then perfused with increasing doses of either SNP (10⁻⁸−10^−1.3 M) or NE (10⁻¹²−10⁻² M) in 5-min increments. SNP, a nitric oxide (NO) donor, induces vasodilation that is endothelium independent and was used to assess vascular smooth muscle-dependent vasodilator sensitivity. NE was used to assess vascular smooth muscle-dependent vasoconstrictor sensitivity. Each agonist that was applied to a BAPN-pretreated site was mixed with 10 mM BAPN to maintain LOX inhibition throughout the protocol.

Time course experiment.

To determine whether the vasoactive effects of LOX inhibition occurred rapidly (in min) or accrued over time (in h), a time course experiment was performed in a subset of five young subjects on a separate day. Two intradermal microdialysis fibers were placed in the ventral forearm. Each fiber was perfused with lactated Ringer solution until the resolution of insertion hyperemia (60–90 min). Sites were randomly assigned either 10⁻⁴ M SNP or 10⁻⁷ M NE. These doses were chosen as they approximated the logEC₅₀ value for each pharmacological agent. Bolus doses of each pharmacological agonist were perfused at 4 μl/min for 5 min. Liquid switches (110, CMA Microdialysis) were used to switch from lactated Ringer solution to agonist perfusion without altering perfusion pressure. After 5 min of agonist perfusion, 10 mM BAPN was perfused through both fibers for 30 min, after which the agonists were then perfused for another 5 min. This process was repeated twice more so that, in total, agonists were perfused once before BAPN treatment and after 30, 60, and 90 min of BAPN treatment.

Data analysis.

Densitometry analysis of Western blots was performed with ImageJ software. Nonmatrix-bound (DOC-soluble) LOX and LOXL2 were normalized to the loading control GAPDH. There is no loading control that can be used to evaluate extracellular matrix-bound (SDS-soluble) LOX/LOXL2; this is because GAPDH and many other cytoplasmic proteins typically used as housekeeping proteins are efficiently solubilized in RIPA lysis buffer. RIPA was used to extract the DOC-soluble portion. Therefore, it is not reliable to use GAPDH to normalize the SDS-soluble portion. Accordingly, extracellular matrix-bound expression data were normalized to the average immunofluorescence of the young or normotensive blots (set equal to 100) from the same set.

Skin blood flow data were stored offline for later analysis. Red blood cell flux was normalized to blood pressure and expressed as cutaneous vascular conductance (CVC; flux/mean arterial pressure). SNP data were expressed as absolute CVC and not normalized to a percentage of maximum CVC, as LOX inhibition was expected to alter maximal vasodilator capacity. LOX inhibition did not change baseline CVC or maximum vasoconstriction in any of the groups (all P > 0.05); therefore, data for NE-mediated vasoconstriction were expressed as a percent change from baseline (20).

Statistics.

Soluble LOX/LOXL2 expression data were compared with one-way ANOVA to compare between the three groups. Extracellular matrix-bound LOX/LOXL2 expression data for young versus hypertensive and normotensive versus hypertensive groups were compared with unpaired t-tests. Pharmacological curve modeling was performed using GraphPad Prism 7.01. Four parameter logistic regressions with variable slopes and no constraints were used to compare logEC₅₀ values. Comparisons were made between logEC₅₀ values with an extra sum-of-squares F-test to determine LOX-mediated changes in vasodilator or vasoconstrictor sensitivity. To assess LOX-mediated changes in the magnitude of vasodilation/vasoconstriction, main effects for treatment (i.e., SNP vs. SNP + BAPN) were assessed with two-way repeated-measures ANOVA with Bonferroni corrections for multiple comparisons when main effects were observed. Time course data were analyzed with one-way ANOVA. Normality for subject characteristic data was evaluated with the Shapiro-Wilk normality test, and subject characteristics were compared with one-way ANOVA. Significance was set to α = 0.05 for all comparisons.

RESULTS

Subject characteristics are shown in Table 1. Young subjects were 20–27 yr old, middle-aged normotensive subjects were 41–62 yr old, and essential hypertensive subjects were 41–63 yr old. Subject age was distributed normally within each group. Measures of resting blood pressure were all higher in hypertensive subjects compared with young and normotensive subjects. Total cholesterol was higher in hypertensive subjects compared with young subjects but was still within the clinically normal range (58a). Ambulatory blood pressure data are shown in Table 2. Almost every measure of ambulatory blood pressure was higher in the hypertensive group compared with the middle-aged normotensive group. The lone exception was nighttime DBP, which was not different between groups but did approach statistical significance (P = 0.053).

Table 1.

Subject characteristics

	Young Subjects	Middle-Aged Subjects	Hypertensive Subjects
Sex (men/women)	6/4	5/5	4/6
Age, yr	24 ± 1	50 ± 2†	53 ± 2†
Body mass index, kg/m2	25.0 ± 1.3	27.3 ± 1.0	27.2 ± 1.4
Systolic blood pressure, mmHg	117 ± 3	110 ± 2	147 ± 2*†
Diastolic blood pressure, mmHg	70 ± 2	71 ± 1	92 ± 3*†
Mean arterial pressure, mmHg	87 ± 2	84 ± 2	112 ± 2*†
Total cholesterol, mg/dl	162 ± 9	183 ± 6	207 ± 9†
LDL-cholesterol, mg/dl	84 ± 6	114 ± 5†	124 ± 8†
HDL-cholesterol, mg/dl	60 ± 5	47 ± 4	65 ± 5*
Blood glucose, mg/dl	88 ± 2	86 ± 1	89 ± 2
HbA1c, %	5.3 ± 0.1	5.5 ± 0.1	5.3 ± 0.1

Values are means ± SE.

^*P < 0.05 compared with middle-aged subjects;

^†P < 0.05 compared with young subjects.

Table 2.

Ambulatory blood pressure data

	Middle-Aged Subjects	Hypertensive Subjects
24-h SBP	110 ± 3	137 ± 2*
24-h DBP	71 ± 1	86 ± 2*
24-h MAP	84 ± 2	103 ± 2*
Daytime SBP	112 ± 3	140 ± 2*
Daytime DBP	72 ± 1	88 ± 2*
Daytime MAP	85 ± 2	105 ± 1*
Nighttime SBP	99 ± 2	117 ± 5*
Nighttime DBP	65 ± 2	74 ± 4
Nighttime MAP	77 ± 2	88 ± 4*

Values are means ± SE. SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure.

^*P < 0.05 compared with middle-aged subjects.

Data for soluble LOX/LOXL2 expression are shown in Fig. 1. There were no differences in soluble LOX expression among the groups (all P > 0.05). Soluble LOXL2 expression was greater in the young group compared with both normotensive (young: 1.60 ± 0.28 and normotensive: 0.85 ± 0.25, P = 0.048) and hypertensive groups (hypertensive: 0.37 ± 0.08, P = 0.0006). There were no differences between normotensive and hypertensive subjects (P = 0.16).

Fig. 1.

A: representative Western blots of soluble-bound lysyl oxidase (LOX)/LOX-like 2 (LOXL2) in two normotensive (N) subjects and two hypertensive (H) subjects. B and C: soluble LOX (B) and LOXL2 (C) expression in young (n = 4), middle-aged normotensive (n = 4), and essential hypertensive (n = 8) subjects. Expression was normalized to GAPDH. Statistics were by one-way ANOVA. *P < 0.05 compared with young subjects.

Expression data for extracellular matrix-bound LOX/LOXL2 are shown in Fig. 2. Matrix-bound LOX expression was higher in the hypertensive group compared with the young group (young: 100 ± 8 and hypertensive: 162 ± 8, P = 0.002), but there was no difference between normotensive and hypertensive groups (normotensive: 100 ± 24 and hypertensive: 159 ± 31, P = 0.17). Matrix-bound LOXL2 expression was greater in the hypertensive group compared with both young (young: 100 ± 19 and hypertensive: 273 ± 44, P = 0.01) and normotensive (normotensive: 100 ± 29 and hypertensive: 661 ± 174, P = 0.01) groups.

Fig. 2.

A: representative Western blots of extracellular matrix-bound lysyl oxidase (LOX) (LOX)/LOX-like 2 (LOXL2) in two normotensive (N) subjects and two hypertensive (H) subjects. B−D: extracellular matrix-bound LOX (B and C) and LOXL2 (D and E) expression in young (n = 4), normotensive (n = 5), and hypertensive subjects (n = 9). Data were normalized as fold changes from the average young or middle-aged normotensive subject analyzed in each set of Western blots. Statistics were by unpaired t-test. *P < 0.05, difference in expression between young/normotensive and hypertensive subjects.

Results from the SNP dose-response protocol are shown in Fig. 3. In young subjects, vasodilator sensitivity to SNP was not altered with LOX inhibition (logEC₅₀ SNP: −4.68 and SNP + BAPN: −3.71, P = 0.11), nor was there an effect of LOX inhibition of the magnitude of SNP-mediated vasodilation (main effect of treatment, P = 0.23). In the middle-aged normotensive group, vasodilator sensitivity to SNP was not altered with LOX inhibition (logEC₅₀ SNP: −3.21 and SNP + BAPN: −3.58, P = 0.78), but there was a significant effect of LOX inhibition (main effect of treatment, P = 0.004), with SNP + BAPN augmenting vasodilation relative to SNP alone. However, post hoc analysis revealed that there were no between-treatment differences in the magnitude of vasodilation at any given dose of SNP within the middle-aged normotensive group (all P > 0.05). In the hypertensive group, LOX inhibition did not alter vasodilator sensitivity to SNP (logEC₅₀ SNP: −2.12 and SNP + BAPN: −1.95, P = 0.99), nor was there a main effect of LOX inhibition on the magnitude of SNP-mediated vasodilation (P = 0.55).

Fig. 3.

Pharmacological curve modeling for the sodium nitroprusside (SNP) dose response with and without lysyl oxidase (LOX) inhibition in young (A), middle-aged normotensive (B), and essential hypertensive (C) groups. LOX inhibition did not alter logEC₅₀ in any group. There was a main effect of LOX inhibition on SNP-mediated vasodilation in the middle-aged normotensive group. Statistics were by repeated-measures ANOVA. *P < 0.05. CVC, cutaneous vascular conductance.

There were between-group differences with the SNP dose response without LOX inhibition. Vasodilation was greater in the young group compared with both normotensive (P < 0.001) and hypertensive (P = 0.004) groups; vasodilation was also greater in the hypertensive group compared with the normotensive group with SNP (P = 0.044). There were no between-group differences with SNP + BAPN treatment.

Data from the NE dose-response protocol are shown in Fig. 4. In the young group, LOX inhibition increased vasoconstrictor sensitivity to NE (logEC₅₀ NE: −6.01 and NE + BAPN: −7.09, P = 0.027). There was a main effect of LOX inhibition in the young group, with LOX inhibition augmenting the magnitude of NE-induced vasoconstriction (P < 0.001); however, post hoc analysis revealed no differences between treatment at any individual dose of NE (all P > 0.05). Vasoconstrictor sensitivity was also increased with LOX inhibition in the middle-aged normotensive group (logEC₅₀ NE: −4.81 and NE + BAPN: −6.98, P = 0.022). There was a main effect of LOX inhibition on the magnitude of NE-induced vasoconstriction in the middle-aged normotensive group, with LOX inhibition augmenting NE-induced vasoconstriction (P = 0.017); however, post hoc analysis revealed no differences between treatment at any individual dose of NE (all P > 0.05). LOX inhibition did not alter sensitivity to NE in the essential hypertensive group (logEC₅₀ NE: −5.98 and NE + BAPN: −6.13, P = 0.79), nor was there a main effect of LOX inhibition on the magnitude of NE-induced vasoconstriction (P = 0.21).

Fig. 4.

Pharmacological curve modeling of norepinephrine (NE)-mediated vasoconstriction with and without lysyl oxidase (LOX) inhibition in young (A), middle-aged normotensive (B), and essential hypertensive (C) groups. Data are expressed as percent changes from baseline (%CVC_baseline). Sensitivity to NE was augmented in the young and middle-aged normotensive groups but not the essential hypertensive group. There was a main effect of LOX inhibition on NE-mediated vasoconstriction in the young and middle-aged normotensive groups. Statistics were by repeated-measures ANOVA. *P < 0.05.

There were between-group differences in the NE dose response without LOX inhibition. Vasoconstriction was greater in the young group compared with the normotensive group (P = 0.0004). There was no difference between young and hypertensive groups (P = 0.18), whereas the difference between normotensive and hypertensive groups approached significance (P = 0.08). With NE + BAPN, vasoconstriction was greater in young versus normotensive subjects (P < 0.001) and young versus hypertensive subjects (P < 0.001), with no difference between normotensive and hypertensive subjects (P = 0.38).

Data from the time course experiment performed on a subset of five young subjects are shown in Fig. 5. Relative to the vasodilator response to a bolus of SNP delivered before LOX inhibition, LOX inhibition did not alter the magnitude of SNP-mediated vasodilation at any time point (all P > 0.05). Similarly, LOX inhibition did not alter the magnitude of NE-mediated vasoconstriction compared with a bolus dose of NE delivered before LOX inhibition at any time point (all P > 0.05).

Fig. 5.

Skin blood flow data in the time course experiment performed in a subset of subjects from the young group (n = 5). Lysyl oxidase inhibition with β-aminopropionitrile (BAPN) did not alter sodium nitroprusside (SNP)-mediated vasodilation (10⁻⁴ M SNP) or norepinephrine (NE)-mediated vasoconstriction (10⁻⁷ M NE) at any time point relative to the BAPN-free baseline. Statistics were by one-way ANOVA. Data are expressed as percent changes from baseline (%CVC_baseline).

DISCUSSION

The main findings from this study were that 1) soluble LOXL2 expression was lower in middle-aged normotensive and essential hypertensive groups relative to the young group; 2) matrix-bound LOX expression was greater in the hypertensive group compared with the young group, and matrix-bound LOXL2 expression was higher in hypertensive group relative to both young and normotensive groups; 3) acute LOX inhibition augmented endothelium-independent vasodilation in the middle-aged normotensive group; and 4) acute LOX inhibition increased vasoconstrictor sensitivity to NE in young and middle-aged normotensive groups but not in the essential hypertensive group. These data suggest that the patterns of LOX and LOXL2 expression change with age and hypertension and that these changes in LOX expression contribute to changes in the role the LOX family plays in regulating cutaneous microvascular function. These findings represent the first-step in vivo human research on the influence of LOX on microvascular function.

Members of the LOX family of enzymes are freely soluble when first secreted from fibrinogenic cells and then become relatively insoluble once associated with the extracellular matrix (26). Soluble LOX mediates the emerging chemotactic and oncogenic actions of LOX, whereas matrix-bound LOX mediates classic collagen cross-linking and vascular integrity roles (26). In animal models of hypertension, members of the LOX family of enzymes are upregulated and vascular stiffness increased (13, 31, 38); however, the potential differences in soluble and insoluble LOX have not been fully elucidated. Our data demonstrate that changes in the expression of soluble and extracellular matrix-bound LOX and LOXL2 occur in hypertensive humans. Extracellular matrix-bound LOX expression was higher in hypertensive subjects compared with young subjects, whereas matrix-bound LOXL2 expression was higher in hypertensive subjects relative to young and normotensive subjects. Furthermore, there were no differences in soluble LOX among the three groups. The higher extracellular matrix-bound LOX and LOXL2 expression likely contribute to the increased collagen cross-linking and vascular stiffening usually observed with hypertension. Furthermore, our findings in the cutaneous microvasculature demonstrate that alterations in LOX family expression with hypertension are not limited to conduit arteries or the heart; future research will determine whether changes in LOX expression with hypertension occur systemically.

The differences in LOX/LOXL2 expression among the young, middle-aged normotensive, and essential hypertensive groups are in line with our functional skin blood flow data. BAPN is an irreversible inhibitor of all LOX isoforms (43) and can elevate soluble collagen (indicative of reduced cross-linking) over a relatively short time period (6 h) (3). However, normal physiological collagen turnover is slow (in wk) (47), equivocating whether LOX/LOXL2-mediated collagen cross-linking would be altered during the acute exposure to BAPN used in this study. We observed greater vasodilation to an endothelium-independent vasodilator and increased sensitivity to a direct smooth muscle vasoconstrictor in the normotensive group. Because of the short timeframe of our study, these vasoactive changes may have been due to inhibition of soluble LOX/LOXL2. In contrast, acute changes in skin blood flow were not observed in the hypertensive group. There are multiple plausible explanations for the lack of a functional change in the hypertensive group. First, the hypertensive subjects likely exhibited multiple pathological adaptations in the microvasculature, including, but not limited to, increased vascular smooth muscle cell stiffness (50), which would limit both vasodilator and vasoconstrictor function. Second, the acute effects of LOX inhibition may be limited in the hypertensive group due to attenuation of soluble LOXL2.

A SNP dose-response protocol was used to evaluate endothelium-independent vasodilation. LOX inhibition did not change vasodilator sensitivity to SNP, as assessed by logEC₅₀, in any subject group. However, there was a main effect for increased absolute vasodilation in response to SNP with LOX inhibition in the middle-aged normotensive group. These data suggest that microvascular stiffness may have decreased via reduced collagen cross-linking in the middle-aged normotensive group. Alterations in vessel responsiveness were not observed in the young group, suggesting that young subjects already have minimal collagen cross-linking without LOX inhibition. There was likely a ceiling effect where LOX inhibition could not further increase smooth muscle responsiveness in the young subjects. Although seemingly counterintuitive, hypertensive subjects may require prolonged LOX inhibition to facilitate changes in pathology-induced upregulated collagen cross-linking due to elevated matrix-bound LOXL2 expression. These results suggest that LOX may be a viable therapeutic target to prevent microvascular remodeling.

A dose response to NE was used to assess vasoconstrictor function. Sensitivity to NE increased with LOX inhibition in both young and middle-aged normotensive groups but not in the essential hypertensive group. There was no change in maximum NE-induced vasoconstriction (10⁻² M) with LOX inhibition in any group. This was contrary to our hypothesis that LOX inhibition would attenuate NE-induced vasoconstriction. The mechanism behind our finding is not immediately clear. A potential explanation involves the effect LOX inhibition has on endogenous production of the vasoactive compoun H₂O₂, which can cause vasodilation (1) or attenuate NE-mediated vasoconstriction (16). H₂O₂ is produced during LOX-catalyzed synthesis of reactive semialdehydes (42) and is physiologically active, having been shown to regulate chemotaxis (30, 41). Hypertensive adults express upregulated provasoconstrictor pathways (2, 23, 48, 54, 55, 61), which likely masks any vasodilatory effect of LOX-derived H₂O₂ in this group. However, this explanation remains largely speculative and will require further investigation.

This study was short term in nature, examining the acute localized effects of LOX inhibition. In animal models, Arem et al. (3) demonstrated that BAPN increased soluble collagen within 6 h; however, soluble collagen was not measured before the 6-h mark. Therefore, it is unknown whether BAPN can mediate alterations in collagen cross-linking on a faster timescale. The data in the present study were collected after ∼2–3 h of administering BAPN, making it difficult to determine whether the results can be attributed to decreased collagen cross-linking. It is likely that inhibition of LOX with BAPN has other non-cross-linking effects that occur over a much shorter time course. Different physiological roles for LOX are emerging, and it has recently been implicated in having angiogenic, chemotactic, and tumor growth properties (4, 14, 26, 27).

To determine whether the vasoactive effects of LOX inhibition occurred immediately or accrued over time, a time course experiment was conducted on a subset of five young participants. No changes in agonist-induced vasoreactivity were observed within the first 90 min of LOX inhibition relative to LOX-intact measures of SNP-mediated vasodilation and NE-mediated vasoconstriction. This suggests that the effects of LOX inhibition begin to manifest around 2–3 h, which concurs with pilot data collected for this study.

A strength of this study was that our approach (intradermal microdialysis) enabled us to safely inhibit LOX in men and women, which has not been previously done. This study helped establish the safety and efficacy of pairing microdialysis and direct LOX inhibition. This will serve as the basis for future projects to more fully delineate the direct in vivo effects of LOX inhibition on human microvascular function and structure. Despite allowing for translation into human subjects, an inherent limitation to this technique is that an index of cutaneous vasodilation had to be utilized as a surrogate measure for stiffness/structure. More direct measures of cutaneous microvascular structure, such as optical coherence tomography, are in their infancy (7). Further development of techniques to assess cutaneous microvascular structure will help determine whether our observations are due to structural changes.

One limitation to this study is that we were not powered to evaluate sex differences. Sex differences in skin blood flow have been observed in some studies (58) but not others (21). We did evaluate all our expression and function skin blood flow data for sex differences; however, none of the results were significant. Future studies should include a larger sample size to elucidate any potential sex differences. Another limitation is that we did not include a group of young adults with hypertension to match with our young normotensive group. However, hypertension in young adults is less common (<10% prevalence) compared with the prevalence in our middle-aged group (∼30–50%) (5). Additionally, the mechanism of the elevation in blood pressure in young humans is typically via an increase in cardiac output, whereas with essential hypertension prevalent with human aging there is more commonly an increase in systemic vascular resistance (36). Our study also would have been aided by measures of LOX and LOXL2 activity. However, because of the subject burden invasive nature of obtaining skin tissue samples, obtaining extra tissue for measures of activity was not feasible in a small pilot study. Future studies should incorporate measures of LOX activity.

Development of specific antagonists for each member of the LOX family would improve our understanding of LOX in the vasculature. At the time of approval for this study, BAPN was the only available LOX inhibitor currently approved by the FDA for investigational use in humans. Only recently has an oral formulation of a LOXL2 inhibitor (PAT-1251) passed a phase I clinical trial. BAPN inhibits LOX and LOXL1–4. Until more specific LOXL antagonists are developed, we cannot determine whether these findings are due to a specific member or a generalized effect of the LOX enzymes.

Conclusion.

We found that matrix-bound LOXL2 expression was higher in essential hypertensive men and women compared with normotensive subjects, which is suggestive of aberrant collagen cross-linking. Conversely, soluble LOXL2 expression was lower in hypertensive subjects compared with middle-aged normotensive subjects. This suggests that alterations in smooth muscle function within the normotensive group were likely due to inhibition of soluble LOXL2. When the LOX family of enzymes was inhibited, endothelium-independent vasodilation was augmented in the middle-aged normotensive group, and smooth muscle-dependent vasoconstrictor sensitivity was augmented in the young and middle-aged normotensive groups. There was no change in smooth muscle function in the essential hypertensive subjects, which was likely due to elevated collagen cross-linking/vessel stiffness attenuating smooth muscle function. These data are important for characterizing LOX/LOXL2 expression and function in men and women and represent a first step for future research on the LOX family of enzymes in humans.

GRANTS

This work was supported by an American College of Sports Medicine Foundation Predoctoral Research Grant (to D. H. Craighead) and National Heart, Lung, and Blood Institute Grant R01-HL093238-07 (to L. M. Alexander).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.H.C., L.S., and L.M.A. conceived and designed research; D.H.C. and H.W. performed experiments; D.H.C. and H.W. analyzed data; D.H.C., H.W., L.S., and L.M.A. interpreted results of experiments; D.H.C. prepared figures; D.H.C. drafted manuscript; D.H.C., L.S., and L.M.A. edited and revised manuscript; D.H.C., H.W., L.S., and L.M.A. approved final version of manuscript.