The Nociceptin/Orphanin FQ Receptor Antagonist UFP-101 Reduces Microvascular Inflammation to Lipopolysaccharide In Vivo

Microvascular inflammation occurs during sepsis and the endogenous opioid-like peptide nociceptin/orphanin FQ (N/OFQ) is known to regulate inflammation. This study aimed to determine the inflammatory role of N/OFQ and its receptor NOP (ORL1) within the microcirculation, along with anti-inflammatory effects of the NOP antagonist UFP-101 (University of Ferrara Peptide-101) in an animal model of sepsis (endotoxemia). Male Wistar rats (220 to 300 g) were administered lipopolysaccharide (LPS) for 24 h (-24 h, 1 mg kg-1; -2 h, 1 mg kg-1 i.v., tail vein). They were then either anesthetised for observation of the mesenteric microcirculation using fluorescent in vivo microscopy, or isolated arterioles (~200 µm) were studied in vitro with pressure myography. 200 nM kg-1 fluorescently labelled N/OFQ (FITC-N/OFQ, i.a., mesenteric artery) bound to specific sites on the microvascular endothelium in vivo, indicating sparse distribution of NOP receptors. In vitro, arterioles (~200 µm) dilated to intraluminal N/OFQ (10-5M) (32.6 + 8.4%) and this response was exaggerated with LPS (62.0 +7.9%, p=0.031). In vivo, LPS induced macromolecular leak of FITC-BSA (0.02 g kg-1 i.v.) (LPS: 95.3 (86.7 to 97.9)%, p=0.043) from post-capillary venules (<40 µm) and increased leukocyte rolling as endotoxemia progressed (p=0.027), both being reduced by 150 nmol kg-1 UFP-101 (i.v., jugular vein). Firstly, the rat mesenteric microcirculation expresses NOP receptors and secondly, NOP function (ability to induce dilation) is enhanced with LPS. UFP-101 also reduced microvascular inflammation to endotoxemia in vivo. Hence inhibition of the microvascular N/OFQ-NOP pathway may have therapeutic potential during sepsis and warrants further investigation.


Introduction
The presence of bacteria in the bloodstream, or sepsis, results in generalised inflammation of blood vessels, hypotension and hypovolemia, which together cause tissue damage and poor perfusion of organs, potentially resulting in multiple organ failure. In the United States there are an estimated 20 million annual cases of severe sepsis [1]. With a mortality rate of 35% this suggests around 700,000 deaths per year, exceeding mortality rates for breast and colon cancer [1].
Pharmacological treatments remain inadequate therefore it is of utmost importance to discover new pathways that could be targeted therapeutically during this potentially fatal condition.
Nociceptin/orphanin FQ (N/OFQ) is an opioid-like peptide and an endogenous ligand for the NOP receptor (previously known as ORL1), which is coupled to the G i /G o signalling pathway [2]. With the exception of kappa;KOP (very weak) there is no activity of N/OFQ at classical opioid receptors (mu;MOP, delta;DOP or KOP [2]). NOP is expressed on aortic endothelium [3], and following intravenous (i.v.) and intracerebroventricular (i.c.v.) administration, N/OFQ caused both hypotension and bradycardia [4][5][6]. N/OFQ also increased aortic blood flow and induced dilation of large blood vessels >200 µm [4,7]. The cardiovascular effects of N/OFQ are absent in mouse knockouts for NOP (NOP -/-) [8]. Within smaller blood vessels (15 to 40 µm) N/OFQ also induces dilation, along with inflammatory responses such as increased permeability and leukocyte-endothelial interactions within postcapillary venules [9]. Furthermore, white blood cells express NOP [10,11], but NOP expression on the microvascular endothelium has not yet been directly identified. Our preliminary experiments using qPCR suggested that protein levels of NOP within the microcirculation may be too low for analysis of expression via this technique. Hence we labelled N/OFQ with a fluorescent marker (fluoroscein isothiocyanate, FITC) to produce FITC-N/ OFQ, which could then be injected locally at high concentrations to identify the distribution of NOP in small vessels.
The N/OFQ-NOP pathway may also play a role in sepsis [12], with plasma N/OFQ concentrations being higher in septic patients who died within 30 days, compared to survivors [11]. Upregulation of N/OFQ has also been reported in rat sensory neurones and hypothalamus in response to lipopolysaccharide (LPS), a component of gram negative bacteria, with antibodies to the receptor for LPS (toll-like receptor-4; TLR-4) also preventing induction of N/OFQ by LPS in murine dorsal root ganglion neurons [13][14][15]. Nevertheless, studies using isolated rat splenocyte cells did not detect significantly increased levels of N/OFQ released from cells until 18 hours of stimulation with LPS [16]. For this reason, combined with the fact that many patients are not treated until sepsis is well established, we studied N/OFQ 20 to 24 hours after the onset of endotoxemia.
Thus far N/OFQ (0.1 mg kg -1 , s.c.) has been found to decrease survival from 30% to 0% at 5 days in the cecal ligation and puncture (CLP) model of sepsis in rats [17]. In conjunction with proposed inflammatory role for N/OFQ [2], we further hypothesised that the function of microvascular NOP-N/OFQ would be altered in an established LPS model of sepsis (endotoxemia) [18,19].
Should the NOP-N/OFQ pathway be upregulated during endotoxemia one would anticipate that manipulating this pathway could be of therapeutic benefit. Only one study has previously utilised the selective NOP antagonist [Nphe 1 , Arg 14 ,Lys 15 ] N/OFQ-NH 2 (UFP-101) [20] during sepsis: in rats, Carvalho et al. demonstrated that 0.03 mg kg -1 UFP-101 improved survival rates from approximately 30% to 65%. 0.03 mg kg -1 UFP-101 also reduced plasma concentrations of TNF-α and IL-1β in response to CLP [17], and when given i.c.v. reduced the hypothalamic pituitary adrenal stress response to LPS [14]. The current study was important to determine whether UFP-101 was similarly effective in the LPS model of sepsis, and whether beneficial effects were reproduced at the microvascular level.
This study aimed to determine the inflammatory role of N/OFQ and its receptor NOP (ORL1) within the microcirculation, along with anti-inflammatory effects of the NOP antagonist UFP-101 (University of Ferfara Peptide-101) in an animal model of sepsis (endotoxemia).

Animals and Housing
Male Wistar rats (200 to 300 g; Harlan, UK, n=47) were obtained from the Field Laboratories at the University of Sheffield, UK. All experiments were performed ethically in accordance with the UK Home Office, Project Licence number 40/2809. Animals were held in the animal facilities for at least 1 week before experimental procedures, exposed to light on a 12: 12 h cycle in a humidity-and temperature-controlled environment, maintained on 0.3% sodium, standard, pelleted, commercial diet and allowed water ad libitum.
Image analysis. Images were monitored using a CCD camera (TK-C13060B, JVC, UK) displayed on a high-resolution monitor (PVM-14N5MDE, Sony, UK) and recorded on to CD-RW using a DataVideo TM CD recorder (VDR-3000, Holdan Ltd, UK) for later off-line computerised image analysis (Capiscope TM , KK Technology, UK) as previously described [19]. Prior to experimentation a 30 minute equilibration period was allowed (T -60 -T -30 ), during which time one area of interest containing an arteriole (15 to 40 µm) and venule (25 to 70 µm) was selected with transmitted light for the purposes of image analysis. At T -30 FITC-BSA (IVM, Study B) also was given via the jugular vein (0.2 g kg -1 ) to allow fluorescent images to be obtained every 10 minutes during experimentation. At each time point the maximum exposure to blue light (460 to 490 nM) was <60 seconds to prevent photoactivation and tissue damage [28].
To measure receptor binding or macromolecular leak Capiscope TM assigned an integer value to the brightness of the fluorescence based on an 8-bit arbitrary gray scale ranging from 0 (black) to 255 (white), with fixed brightness and contrast levels. Receptor binding (IVM, Study A) was assessed in each animal by using Capiscope to place a box at three sites where 'hot-spots' of receptor binding could be seen. Macromolecular leak (Study B) was measured at three points within the interstitium adjacent (<2 mm) to a randomly selected postcapillary venule, Rolling leukocytes were assessed within a relatively straight length of vessel, using transmitted light, by counting the numbers that passed a fixed point over 30 seconds [9].
Experimental groups. In study A the mesenteric artery was also cannulated to administer FITC-N/OFQ (200 nM kg -1 , 40 nM) intra-arterially at T 0 (n=5). Images of the microcirculation were then recorded continuously for 60 seconds, which was the duration of visible fluorescence. At the following time point (T -15 ), unlabelled N/OFQ (200 nM kg -1 , 40nM) was administered 15 seconds before 200 nM kg -1 FITC-N/OFQ in order to occupy the NOP receptor site.
In study B animals received LPS (Serotype B5: 055) or the equivalent volume of saline (0.1ml 100g -1 ) into the tail vein at -24 h (1 mg kg -1 ) and -2 h (0.5 mg kg -1 ) prior to experimentation (T 0 ) in (i) control, n=6; (ii) LPS, n=6, (iii) UFP-101, n=6 and (iv) LPS + UFP-101, n=6 groups. Animals received UFP-101 (150 nM kg -1 , 62.5nM), or the equivalent volume of saline (1 ml kg -1 ) into the jugular vein at the baseline recording (T 0 ). Images of the microcirculation were recorded for 60 seconds with transmitted light and 30 seconds with blue light (460-490nM) every 10 minutes at T 10 , T 20 , T 30 and T 40 . A full response curve was thus constructed, but data are shown graphically only at T 40 for ease of interpretation as similar patterns of response were observed at earlier time points.
At the end of the procedure rats were killed humanely in accordance with UK Home Office procedure using an overdose of anesthetic followed by cervical dislocation.

Pressure myography
Surgery. Rats were killed in accordance with UK Home Office requirements involving cervical dislocation. This was followed by rapid removal of the ileum and adjacent mesentery from either control or LPS-treated animals. The mesentery was placed in HEPES-PBS solution at 4°C and a third order artery (~200µm) dissected free from surrounding adipose tissue. Vessels were carefully transferred to the organ bath and mounted on glass cannulae at 60mmHg (Living Systems Instrumentation, Burlington, Vermont, USA). Vessels were stabilised and maintained as previously described, with the Living Systems video dimension analyser used to determine luminal diameters 21 . In previous studies it has been determined that rat mesenteric vessels do not respond to abluminal N/ OFQ, hence N/OFQ was administered intraluminally using predetermined doses [9].
After performing a concentration response curve with U46619 (10 -10 to 10 -6 M), all vessels were pre-constricted with an EC 80 of U46619 (2x10 -7 M). U46619 is synthetic analogue of prostaglandin H 2 which acts as a thromboxane A 2 receptor agonist and commonly used in myography to pre-constrict isolated vessels before studying vasodilator responses. Thus following pre-constriction at a constant pressure (60 mmHg) HEPES (i, iii) or the calculated EC 80 dose of N/OFQ (10 -5 M) (ii) was then allowed to flow through the lumen of pre-constricted vessels for 20 minutes at a flow rate of 15 µl min -1 [9]. Measurements of luminal diameter were taken every minute, but data are shown such that the percentage change in the diameters of vessels is reported between T 0 and T 20 . At the end of the protocol, the organ bath was washed out and acetylcholine (ACh, 10 -5 M) was added to pre-constricted vessels to ensure that endothelial integrity had been maintained.

Statistics
Parametric data are presented as mean (+SEM), with statistical analysis performed using two-way ANOVA or student t-test. However, all in vivo data were non-parametric and presented as median and interquartile ranges (25 th to 75 th ), with n = 5 or 6 animals used in each group, as determined using preliminary data. Data are expressed as box and whisker plots, except where negative values were present within data and this was not possible. In study A paired data were analysed using the Wilcoxon paired signed rank test at equivalent time points. In study B, if Kruskall-Wallis test identified significant differences between groups, analyses was performed using the Mann-Whitney U-test. Likewise, comparing data to baseline, if the Freidman test identified a significant difference then the Wilcoxon signed ranks test was performed. To avoid significance being obtained due to multiple tests, only one time point underwent statistical evaluation using post-hoc tests. All statistical analysis was undertaken by an independent statistician using SPSS version 16.0. Results were considered statistically significant at p < 0.05.

Preliminary characterisation of FITC-N/OFQ
In receptor binding experiments FITC-N/OFQ showed high affinity (pK i 9.60 compared to 9.85 for N/OFQ) for NOP and at least 4000 fold selectivity over KOP ( Figure 1A). As anticipated the high affinity binding of FITC-N/OFQ resulted in a high potency stimulation of the binding of GTPγ[ 35 S] to membranes prepared from cells expressing NOP ( Figure 1B). There was no difference in potency (pEC 50 ~8.4) or efficacy (E max : 4.7 to 5.2 expressed as stimulation factor) between the native peptide and FITC-N/OFQ. Based on this simple in vitro characterisation FITC-N/OFQ does not differ substantially from native N/OFQ in terms of affinity, selectivity, functional potency and efficacy indicating its suitability for use in the in vivo experiments described below.

Evidence for N/OFQ specific binding within mesenteric microcirculation
In vivo, FITC-N/OFQ bound to specific sites on the endothelium of arterioles, indicating sparse distribution of NOP receptors in the mesentery in control conditions (Figure 2). This binding appeared to be specific, as N/OFQ prior to FITC-N/OFQ (competing for the same receptor site) reduced the the ability to bind at these sites (Figures 2 and 3).

Cardiovascular and microvascular effects of N/OFQ
In vivo, dose dependent hypotension, dilation of mesenteric arterioles and venules, along with macromolecular leak were observed with N/OFQ in agreement with previous studies (data not presented) 9 . In addition, pressure myography demonstrated that N/OFQ caused isolated arterioles to dilate by 32.6 +8.4%, compared to only 6.8 +3.8% in controls (p=0.031).

Discussion
This study has demonstrated that UFP-101 exhibits antiinflammatory properties within the microcirculation, specifically it reduced macromolecular leak and leukocyte rolling, this potentially being an important contributor to reduced mortality in septic animals treated with UFP-101 17 . Increased blood volume is the basis of current therapeutic strategies, and achieving this via reducing microvessel permeability, in turn increasing blood flow and organ perfusion, confirms promising therapeutic potential for NOP antagonists.
Using a range of complementary experiments and a novel fluorescent analogue of N/OFQ our data also suggested that NOP receptors are expressed on small foci of microvascular endothelium of rat mesentery. This is in agreement with expression of NOP previously reported on the endothelium of larger blood vessels, namely rat aorta [3]. Nevertheless, the present data suggest that expression may be quite sparse in these small vessels. This has functional consequences for NOP partial agonists as they may then behave as antagonists [29]. In conjunction with previous studies, identification of NOP receptors within the microcirculation was an important foundation for understanding how UFP-101 exerted beneficial effects during endotoxemia.
Perhaps the most important part of the study was the therapeutic potential of UFP-101 in an LPS model of sepsis. In agreement with previous observations, that UFP-101 was antiinflammatory during CLP [17], we also determined that UFP-101 reduced macromolecular leak and reduced leukocyte rolling at the later time point in response to LPS. Furthermore, UFP-101 reduced vasodilation of venules in response to LPS, often associated with excessive production of pro-inflammatory cytokines. Effects on arterioles, which regulate blood pressure, were not observed, but mesenteric venules did respond to LPS in agreement with our previous studies at 0 to 4 h [30]. This may be a compensatory mechanism to preserve blood flow to a vital organ. In conjunction with known reductions in proinflammatory cytokines, the ability of UFP-101 to reduce microvascular inflammation may relate to the beneficial effects on mortality during CLP [17]. In agreement, there are increased plasma levels of N/OFQ in patients who die from sepsis versus those who do not [31]. Moreover in a very recent study tracking the time course of sepsis in patients on the intensive care unit we have data demonstrating an increase in plasma N/OFQ levels that falls when the patients recover (in this context patients are their own controls) (Thompson, Serrano-Gomez, McDonald, Ladak, Bowrey and Lambert, 2013, unpublished). Of note however, our animal model of sepsis was not particularly severe, as all animals survived the duration of the experiment with hypotension <20%. Whilst non-lethal endotoxin models of sepsis reflect more accurately the insidious onset of the disease, ultimately septicemia is a disease involving the entry of live bacteria, such as E. Coli into the bloodstream. In CLP models which reflect this, mortality rates are high: for example the study by Carvalho et al. demonstrated a 70% mortality rate [17]. It is thus re-assuring that these present data with LPS are in agreement with this study. With i.v. administration of UFP-101 we also discovered no reversal of the tachycardia or mild hypotension induced by LPS. This finding is in agreement with direct NOP-N/OFQ inflammatory effects on the microcirculation, but not blood pressure, being a locally mediated mechanism involving histamine [9].
In addition to macromolecular leak, dilation of isolated arterioles (~200µm), was certainly increased by exposure to LPS. This suggests that the function of NOP is enhanced during exposure to LPS; further studies are required to confirm this in smaller vessels in vivo and determine the mechanism by which LPS increases NOP expression and/or intracellular signalling. We could not repeat the FITC-N/OFQ experiments in LPS-treated animals, as macromolecular leak interfered with the fluorescent measurements. Interestingly our initial attempts with immunohistochemistry to reproduce the study performed by Granata et al. on the aortic endothelium were unsuccessful using sections of mesenteric microcirculation [3]. This may be because expression is sparse ( Figure 2) and radioligand binding studies with [ 3 H] UFP-101 are complicated by low radioligand specific activity [32].
To develop this point further, a major criticism of our work might be that we have not demonstrated the presence of NOP using more conventional techniques; (i) radioligand binding and (ii) Western blotting. Due to the very small amount of tissue that can be harvested from the vasculature and the presumed ultra-low expression it was not possible to perform a standard binding experiment of the type shown in Figure 1, even if we employed a high specific activity radioligand ([ 125 I] N/OFQ) [33]. The amount of tissue may have been sufficient for Western blotting but, as with most members of the opioid family, there are no selective antibodies for use [34]. In the absence of these, our functional data provides strong evidence for functional NOP and, to reiterate, the completely novel approach used in our present work (FITC-N/OFQ) demonstrated sparse expression in the microvasculature which would not have been possible by other means.
One further limitation of this study is that control animals may have experienced a degree of unavoidable inflammation at baseline, due to the surgery required for observation of the microcirculation using intravital microscopy. In addition, our cardiovascular (and microvascular variables) differed at baseline in all treated groups because animals were already 'septic'. Indeed, an important limitation of studies using anesthesia/surgery and IVM to observe the mesenteric microcirculation, is that it is not possible to integrate a nonseptic baseline or an unanesthetised control. Nevertheless, if there was some degree of inflammation at baseline due to surgery, these data demonstrated that it was reduced in the presence of UFP-101.
Despite the interesting data with inflammatory markers, we could not make firm conclusions as to whether intravenously administered UFP-101 given to non-septic animals truly caused microvascular constriction, as we are uncertain whether this observation was due to differences in the diameter of vessels at baseline or an immediate effect of UFP-101. Despite this, as part unpublished data within previous studies we also observed vasoconstriction with UFP-101 [9]. Compared to larger blood vessels the mesenteric microcirculation as studied here (<40 µm) receives little neural control [35]. Thus, regardless of the baseline status vascular diameters a NOP dependent (rather than neural) mechanism must be engaged for constriction to occur via UFP-101. Taken together with our data utilising isolated vessels and N/OFQ, it is possible that NOP receptors may regulate vessel diameters via endothelial as well as central/neural control mechanisms, as we have alluded to in previous studies proposing an endothelial NOP-histamine pathway [9].
In this study intra-arterially administered UFP-101 had no effect on MAP, similar to studies whereby UFP-101 was given i.c.v. to mice [36]. Further studies likewise reported no differences in MAP between NOP -/-and wild-type mice [8]. However, our findings of tachycardia are in contrast to Burmeister & Kapusta [36]. Nevertheless, similar to our data, 300 pmol UFP-101 injected into the paraventricular nucleus (PVN) of the hypothalamus caused increased heart rate and renal sympathetic nerve activity in conscious Sprague-Dawley rats, along with a non-significant increase in MAP [37]. In our previous studies we did not report the effects of UFP-101 over time, rather its effects on cardiovascular response to N/OFQ [9]. Thus in conjunction with the literature, our new evidence with UFP-101 suggests that the vascular mechanisms of NOP are multifactorial, differing between species and the organ of interest. N/OFQ may also have opposing effects depending upon the route via which it is administered.
At a cellular level activation of NOP will reduce cAMP formation, close voltage gated Ca 2+ channels and open K + channels to produce hyperpolarisation. In a neurone this combination of events is easy to collate into an inhibition of neurotransmission. In addition to these classical signalling events there is also evidence of increased MAP kinase activity [2]. We do not know how these events transfer to the vascular endothelium, but a modulation of the driving force for dilation (increased Ca 2+ , NO activation or cAMP generation) would seem likely targets and these will need rigorous experimental verification in further studies.
In conclusion, we have evidence that the endothelium of the rat mesenteric microcirculation expresses NOP receptors. Furthermore, in agreement with previous studies using a cecal ligation model, UFP-101, a selective antagonist of NOP, reduces microvascular inflammation during an LPS model of sepsis in rats. Animal studies with small data sets should be considered as preliminary data and thus clinical studies would now be useful to determine if this potentially beneficial therapeutic effect could translate into humans.