Modulation of vascular contraction via soluble guanylate cyclase signaling in a novel ex vivo method using rat precision‐cut liver slices

Abstract Fibrotic processes in the liver of non‐alcoholic steatohepatitis (NASH) patients cause microcirculatory dysfunction in the organ which increases blood vessel resistance and causes portal hypertension. Assessing blood vessel function in the liver is challenging, necessitating the development of novel methods in normal and fibrotic tissue that allow for drug screening and translation toward pre‐clinical settings. Cultures of precision cut liver slices (PCLS) from normal and fibrotic rat livers were used for blood vessel function analysis. Live recording of vessel diameter was used to assess the response to endothelin‐1, serotonin and soluble guanylate cyclase (sGC) activation. A cascade of contraction and relaxation events in response to serotonin, endothelin‐1, Ketanserin and sGC activity could be established using vessel diameter analysis of rat PCLS. Both the sGC activator BI 703704 and the sGC stimulator Riociguat prevented serotonin‐induced contraction in PCLS from naive rats. By contrast, PCLS cultures from the rat CCl4 NASH model were only responsive to the sGC activator, thus establishing that the sGC enzyme is rendered non‐responsive to nitric oxide under oxidative stress found in fibrotic livers. The role of the sGC pathway for vessel relaxation of fibrotic liver tissue was identified in our model. The obtained data shows that the inhibitory capacities on vessel contraction of sGC compounds can be translated to published preclinical data. Altogether, this novel ex vivo PCLS method allows for the differentiation of drug candidates and the translation of therapeutic approaches towards the clinical use.


| INTRODUC TI ON
Nonalcoholic fatty liver disease (NAFLD) encompasses a disease spectrum from excessive hepatic fat accumulation to chronic liver inflammation (nonalcoholic steatohepatitis (NASH)) and around 24% of the population in Europa has NAFLD. 1 Accumulation of triglycerides in hepatocytes as seen in NAFLD 2 leads to initiation of inflammation and fibrosis. Reactive oxygen species (ROS) as part of the inflammatory response induce vascular remodeling, fibrosis, and wound healing processes 3 which drive NALFD progression to NASH. 3 Fibrosis can eventually lead to liver cirrhosis, a condition where there is loss of cell organization leading to reduced liver function, 3 microcirculatory dysfunction, 4 and alterations in the blood vessel architecture.
These structural changes generate elevated pressure in the portal vein 5 in which elevated levels of vasoconstrictors such as endothelin-1 and serotonin contribute. 6,7 Currently, there is no treatment for portal hypertension. 8 Evaluation and differentiation of new drug candidates for the treatment of portal hypertension require a translatable model.
A widely used and direct in vivo measurement of portal pressure is achieved by catheterization of the portal vein 9 and analysis of the hepatic venous pressure gradient (HVPG). 9 However, this direct measurement is invasive, difficult, and allows only a low throughput.
In the present study, we established an ex vivo method based on precision-cut liver slices (PCLS) to directly assess blood vessel dilation and contraction in liver explants.
Contraction has been induced in our ex vivo method by two potent modulators of liver vascular tone, serotonin (5HT) and endothelin-1 (ET-1). 10,11 ET-1 is increased in NASH patients, serum levels correlate with the severity of fibrosis, and blocking ET-1 decreases fibrosis making it of interest when investigating therapeutic concepts for NASH. 12 The beneficial effect of targeting the serotonin pathway is already established with the use of a 5-HT 3 receptor antagonist which could ameliorate steatosis and inflammation in the liver of ob/ob mice. 13 A potential therapeutic target to enhance vasodilation and thereby reduce portal hypertension is soluble guanylate cyclase (sGC), an enzyme acting as a receptor for the vasodilator nitric oxide (NO).
Binding of NO to sGC leads to the generation of cyclic guanosine monophosphate (cGMP), an important second messenger that induces vascular smooth muscle cell relaxation and blood vessel dilation. 14,15 The functional sGC enzyme consists of an α-subunit (α1 or α2) and a β-subunit (β1 or β2) containing a haem prosthetic group which is responsible for NO-sensing when the iron is in a reduced state. A conformation change induced by NO binding to the ferrous iron results in stimulation of catalytic activity while oxidation of the haem under conditions of oxidative stress renders sGC insensitive to NO.
Increased enzymatic activity of the sGC enzyme can be achieved by either sGC stimulators which increase the sensitivity of the haem prosthetic group for NO or sGC activators that function independently of NO on haem-oxidized or haem-free sGC enzyme. 16 Oxidation of the haem under conditions of oxidative stress as observed in NASH patients has the potential to render the sGC enzyme insensitive to NO 17 making an sGC activator potentially more effective than an sGC stimulator. For the comparison of sGC stimulators and activators, we used published representative compounds.
Riociquat, an sGC stimulator, has shown efficacy as a vasodilator in a phase III trial for pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension 18 and as a treatment for portal hypertension. The compound BI 703704, known to modulate sGC by dose-dependently increasing cGMP levels independent of the oxidative status of the enzyme, showed in vivo potency in the treatment of diabetic nephropathy and portal hypertension. 19,20 We sought to explore the translation of the beneficial role of sGC in treating portal hypertension in vivo and its link to serotonininduced blood vessel contraction in a new ex vivo model. PCLS have been efficiently used for the characterization of fibrotic pathways and the effect of compound treatments. 21,22 Here we established their use for the investigation of compound-related effects on serotonin-induced vessel contraction using Ketanserin and sGCrelated compounds. Both an sGC activator and an sGC stimulator were able to induce a dose-dependent inhibition of blood vessel contraction using PCLS from normal mice. Using fibrotic liver tissue,

Summary
What is already know?
• Precision cut liver slices conserve hepatic physiology and response to fibrotic stimuli • Limiting methods to analyze blood vessel function available, raising the need for new translatable methods • Portal hypertension as a response to fibrotic processes in NASH patients.

What did this study adds?
• A description of a pharmacological treatment and analysis of hepatic vessel function under fibrotic conditions.
• Demonstration of the sGC activator induced vasodilation under oxidative stress conditions in the liver.

Clinical significance
• The ex vivo method allows for pre-clinical to clinical translation of pharmacodynamic effects.
the sGC activator showed a superior effect in reducing contraction under diseased conditions with elevated oxidative stress.
Overall, the PCLS model system allows for the analysis of vessel diameter and the prioritization of potential therapeutics for portal hypertension and the translation to more complex in vivo models. 20 2 | MATERIAL S AND ME THODS

| Materials
Coring press was obtained from Alabama Research & Development.

| Animals
Male Wistar rats (300-400 g) from Charles River Research Models and Services Germany GmbH were used for all control experiments.
Male rats for selected for this study based on the know differences observed between the severity of liver damage induced by CCl4 in male and female rats. 23

| Preparation of PCLS
Tissue cores with a diameter of 8 mm were punched from rat livers

| Histology
After the contraction experiment, slices were fixed in 4% paraformaldehyde or directly stained. Slices were placed in incubation buffer (1× HBSS, 8.9 mM sodium bicarbonate, 5% fetal bovine serum, Sigma Aldrich) containing 5 µg/ml Hoechst for 30 min shaking at F I G U R E 1 Ex vivo model for measuring blood vessel contraction in precisioncut liver slices (PCLS). Tissue cores were punched out from rat livers. PCLS were prepared by means of a Krumdieck tissue slicer. Each slice was incubated at 37°C under high oxygen and shaking conditions. The day after slicing medium was changed. Slices containing blood vessels were selected and incubated with 1 µM or 10 µM compound for 30 min at 37°C. PCLS were transferred into cavities of standard multi-well cell culture plates, mounted by a platinum ring to prevent floating and subjected to videomicroscopy. Serotonin was added to PCLS and pictures were taken every 10 s for 7 min. Blood vessel area was analyzed by Axiovision Software. H&E and Hoechst staining of PCLS showed the presence of the portal vein and the central vein in the slices 4°C followed by a washing step of 20 min in fresh incubation buffer.
For analysis, slices were placed on an object slide and covered using a mounting medium. Hematoxylin and eosin (H&E) staining was performed on liver slices (3 µm thickness) embedded in paraffin. Slides were scanned with a Zeiss Axio Scan.Z1.

| Viability assay: ATP
Directly after slicing and after 24 h of incubation, PCLS were collected for ATP measurements in Precellys ® Keramik-Kit 1 mm tubes in 1 ml sonification buffer containing 70% ethanol and 2 mM EDTA (Sigma Aldrich). Slices at 0 h were collected at 4°C, which is the temperature during slicing, but slices at 24 h are collected at 37°C. Homogenization of PCLS was achieved with the use of the FastPrep System from MP Biomedicals and followed by centrifugation (18000g, 5 min, 4°C). ATP levels were measured by ATP Bioluminescence Assay Kit (Roche) in the supernatants while the pellets were used for protein determination. The pellets were dried and homogenized in 1 M NaOH. Protein amounts were determined by Pierce™ protein assay kit (Thermo Scientific).

| RT-PCR
PCLS were stored at −20°C in Precellys ® Keramik-Kit 1.4 mm tubes containing 400 µl RLT (Qiagen) +1% β-mercaptoethanol. Slices were lysed using the Precellys Evolution twice at 2600 g for 30 s each. 400 µl of 70% ethanol was added before mRNA isolation according to the RNAeasy protocol from Qiagen. Following the manufacturer protocol, 1 µg RNA was transcribed into cDNA using High capacity cDNA Archive Kit from Applied Biosystems (#4322169).
The cDNA was amplified in a quantitative real-time PCR with gene-specific primers from Applied Biosystems. The following

| Data and statistical analysis
The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology. 27 Group size was based on a predictive effect size of 15% contraction in our control group with a mean of 100% and SD of 8%. A group size of n > 5 is needed with α: 0.05 and β: 0.80.
At the end of each contraction experiment, the contractile response of each slice was determined by the addition of ET-1. Due to this stringent quality control, which was applied after the samples were taken, several samples did not reach the final analysis. In accordance with this high internal standardization and quality control for the vessel function, the experiments allow for the reduced number of samples.
Another deviation concerning group size is seen in the DMSO control group in the contraction experiment, where more than n:5 were included. As different rats were used, we put a high value to test multiple DMSO controls for each rat to obtain a solid characterization of each rat using the new setup before testing compound incubation on the other slices for each rat. DMSO control was taken at different times of the day to overcome the variation timing possibly can cause.   In the first step, the viability of the slices was assessed and confirmed by ATP measurements (Figure 2A)

| Both the sGC stimulator Riociguat and the activator BI 703704 prevent serotonin-induced contractions in blood vessels in healthy PCLS
To investigate the role of the sGC pathway in the contraction of blood vessels, PCLS were pre-incubated with 10 µM Riociguat (sGC stimulator) or 10 µM BI 703704 (sGC activator). 19 The contraction induced by serotonin was completely blocked by BI 703704 (Figure 3C and D) early after the addition of serotonin.
In contrast, Riociguat was not able to fully inhibit vessel contraction in PCLS and there was a delayed response compared to BI 703704 treatment ( Figure 3C and D).

| PCLS from CCl 4 -treated animals show changes in sGC subunit expression, oxidative stress response, and fibrosis signatures
To determine if the difference between the sGC stimulator and sGC activator is more evident under oxidative stress conditions, PCLS from rats treated with CCl 4 were prepared. Sprague Dawley rats were used for our CCl 4 model as these rats are more prone to a cirrhotic phenotype compared to Wistar rats. 29 The expression of all four sGC subunits was analyzed in PCLS from vehicle and CCl 4 -treated animals ( Figure 4). In PCLS from the

(D)
Riociguat or BI 703704 did not alter the expression of the four subunits ( Figure 4C-F).

CCl 4 was used as a model for NASH in which oxidative stress is
increased. Reactive-oxygen species (ROS) are neutralized by protective mechanisms 30 which are induced by the transcription factor Nrf2: These include haem oxygenase-1 (HMOX-1) and NAD(P)H:quinone dehydrogenase 1 (NQO1). 30 We investigated the mRNA expression of certain oxidative stress markers in PCLS ( Figure 5). Figure 5A shows that NQO1 is increased in PCLS from CCl 4 -treated rats compared to vehicle-treated rats at 0 h. After incubation for 24 h HMOX1 was increased in PCLS from vehicle and CCl 4 -treated animals ( Figure 5B).

| Blood vessels in PCLS obtained from CCl 4 -treated animals showed a reduced contractile response to serotonin
CCl 4 was used as a NASH disease model, inducing a fibrotic phenotype responsible for alterations in the blood vessel architecture. PCLS obtained from CCl 4 -treated animals showed a remarked reduction in the amount of serotonin-induced contraction. CCl 4 -treated animals only had 5% contractility, while 16% contraction was seen in PCLS from healthy animals ( Figure 6A-D)). Pre-incubation with 10 µM Ketanserin was able to block the serotonin-induced contraction ( Figure 6B). Riociguat or BI 703704, prevented serotonin-induced contraction in PCLS of healthy animals ( Figure 6A and B). Riociguat and BI 703704 were also able to prevent the 5% vessel contraction induced by serotonin in PCLS from CCl 4 animals ( Figure 6C and D).

| Only the sGC activator induces blood vessel relaxation in PCLS obtained from CCl 4 animals
To assess the relaxation potential of the sGC compounds under fibrotic conditions, PCLS from CCl 4 -treated rats were analyzed.
Relaxation was measured as an increase in blood vessel area com-

| D ISCUSS I ON AND CON CLUS I ON
In NASH patients the hepatic vascular resistance is increased due to structural changes in the liver related to fibrosis. An increased resistance as observed in fibrotic or cirrhotic patients is additionally due to an increase in contractile agents such as endothelin-1, serotonin, and simultaneously a diminished availability of the relaxant agent NO. 6,31 Both endothelin-1 and serotonin induce a contractile response and are known to regulate blood flow in both the portal and sinusoidal vessels and increase the portal pressure in rats. 11 In cirrhotic patients, higher plasma levels of serotonin are observed and inhibition leads to a reduced portal pressure in these patients. 6 These literature data provide evidence for the importance of serotonin in cirrhotic portal hypertension and that modulating the serotonin pathway might be beneficial in the treatment of portal hypertension.
Our new ex vivo model based on PCLS (Figure 1) showed that Ketanserin, a vasodilator, prevents liver blood vessel contraction induced by serotonin (Figures 2 and 8 ing. The effect of serotonin seems to be similar in the portal and central vein, so the amount of contraction cannot be used as an identification method for the different veins. 11,34 One way to optimize the identification is using a higher resolution microscope to differentiate between the different veins without staining. Other constraints concern the viability of the slice which can only be assessed after pre-incubation therefore multiple slices are needed to allow for a sufficient group size at the end. Lastly, blood vessels in PCLS from severely diseased animals have the tendency to be closed, increasing the difficulty in performing contraction experiments. Overall, this model based on PCLS is suited to test the effect of contractile agents, such as serotonin and ET-1 (Figure 2), and to test the inhibitory properties of compounds, which makes the model suitable for analyzing therapeutic concepts, such as portal hypertension. The use of the model for drug discovery as shown here is based on the link between serotonin and the NO/sGC/cGMP pathway. We demonstrated that both sGC compounds can prevent serotonininduced contraction in PCLS from naïve rats (Figures 3 and 8). This result is consistent with the results obtained using Riociguat in in vivo models. In a bile duct ligation (BDL) study, Riociguat was able to reduce portal pressure in rats 35 indicating that the ex vivo method described can be predictive for in vivo experiments (Figure 3). Based on the data presented by Brusilovskaya et al, we believe that a PDEinhibitor like Tadalafil, which inhibits portal pressure in a BDL rat model could also inhibit blood vessel contraction as seen with our PCLS method. 20 However, we speculate that targeting the upstream components such as sGC might induce a greater effect than targeting the downstream inhibition of cGMP degradation by PDE inhibition. We hypothesized that the effect of the sGC activator will become more evident in the presence of oxidative stress. Therefore, In Figure 6 we demonstrate that the contractile effect of serotonin was almost absent in liver slices from CCl 4 -treated animals ( Figure 6A and C), which could point to an increased level of vasoconstriction due to liver injury. 38 This is consistent with the findings of Ruddell and colleagues that the hepatic artery only shows a minor increase in serotonin-induced contraction in cirrhotic patients compared to healthy patients. 11 The plasma of patients with liver cirrhosis and portal hypertension showed an 18-fold increase of serotonin making the blood vessel less sensitive to exogenously added serotonin. 11 To analyze the effect of both the sGC activator and the sGC stimulator on relaxation, we used a different setup, in which PCLS were incubated with the compound directly and monitored by videomicroscopy. We were unable to induce relaxation with any of the compounds in PCLS of the vehicle-treated rats ( Figure 7A and B), potential due to maximal relaxation in these vessels. However, the sGC activator BI 703704, but not the sGC stimulator Riociguat, was able to induce relaxation in PCLS of CCl 4 -treated animals ( Figure 7C and D). This supports the hypothesis of an sGC activator having beneficial effects under conditions of oxidative stress compared to an sGC stimulator.
In conclusion, the ex vivo method based on PCLS described here captures aspects of the NASH phenotype as seen in humans and can be used to study blood vessel contractility in the liver, evaluate potential therapeutic agents as shown for the sGC pathway (Figure 8)

F I G U R E 8
Overview of the effect of an sGC stimulator and sGC activator on blood vessel contraction and allow for translation of therapeutically findings to pre-clinical animals models.

| DISCL SOURE
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

E TH I C S A PPROVA L
The animal studies were conducted according to approved animal licenses of Boehringer-Ingelheim and conform to national guidelines for animal welfare.

ACK N OWLED G M ENTS
We would like to acknowledge Lars Pischzan, Katja Thode and Anke Voigt for their excellent technical support. My grateful thanks are also extended to Dr. Chris Sarko and Dr. Todd Bosanac for the design of the compound BI703704.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within the article [and/or] its supplementary materials. The data that support the findings of this study are available from the corresponding author upon reasonable request.