Quantitative real-time measurement of endothelin-1-induced contraction in single non-activated hepatic stellate cells

Although quiescent hepatic stellate cells (HSCs) have been suggested to regulate hepatic blood flow, there is no direct evidence that quiescent HSCs display contractile abilities. Here, we developed a new method to quantitatively measure the contraction of single isolated HSCs and evaluated whether endothelin-1 (ET-1) induced contraction of HSCs in a non-activated state. HSCs isolated from mice were seeded on collagen gel containing fluorescent beads. The beads around a single HSC were observed gravitating toward the cell upon contraction. By recording the movement of each bead by fluorescent microscopy, the real-time contraction of HSCs was quantitatively evaluated. ET-1 induced a slow contraction of non-activated HSCs, which was inhibited by the non-muscle myosin II inhibitor blebbistatin, the calmodulin inhibitor W-7, and the ETA receptor antagonist ambrisentan. ET-1-induced contraction was also largely reduced in Ca2+-free conditions, but sustained contraction still remained. The tonic contraction was further diminished by the Rho-kinase inhibitor H-1152. The mRNA expression of P/Q-type voltage-dependent Ca2+ channels (VDCC), as well as STIM and Orai, constituents of store-operated channels (SOCs), was observed in mouse non-activated HSCs. ET-1-induced contraction was not affected by amlodipine, a VDCC blocker, whereas it was partly reduced by Gd3+ and amiloride, non-selective cation channel blockers. However, neither YM-58483 nor SKF-96365, which inhibit SOCs, had any effects on the contraction. These results suggest that ET-1 leads to Ca2+-influx through cation channels other than SOCs and produces myosin II-mediated contraction of non-activated HSCs via ETA receptors, as well as via mechanisms involving Ca2+-calmodulin and Rho kinase.


Introduction
Hepatic stellate cells (HSCs), located in the space of Disse between sinusoidal endothelial cells and hepatocytes, constitute 5%-8% of the total number of liver-resident cells [1]. Under physiological conditions, HSCs exist in a quiescent state, called quiescent HSCs (qHSCs), accumulating lipid droplets containing vitamin A [2]. Upon liver injury by alcohol consumption or viral infection, qHSCs are activated and transdifferentiated into myofibroblast-like activated HSCs (aHSCs) that produce collagen and undergo enhanced cell proliferation. Accordingly, the activation of HSCs is likely a cause of liver fibrosis and cirrhosis. In addition, aHSCs express α-smooth muscle actin (α-SMA), smooth muscle myosin, and L-type voltagedependent Ca 2+ channel (VDCC) [3], thereupon exhibiting contractile properties [4]. aHSCs are also suggested to be a cause of portal hypertension [4]. In contrast, the contractile properties of qHSCs under physiological conditions are not clear. qHSCs encircling the sinusoid are anatomically similar to pericytes in capillaries. Therefore, qHSCs are postulated to participate in the regulation of blood flow in the liver [5]. Moreover, the contraction of rat qHSCs has been suggested to contribute to sinusoidal resistance [6]. Interestingly, endothelin-1 (ET-1), which was originally identified as a 21 residue potent vasoconstrictor peptide [7], has been shown to decrease sinusoidal diameter and increase sinusoidal resistance in perfused rat livers [8]. Isolated rat HSCs have been shown to display contractile activity in response to ET-1 [9]. The contraction of human HSCs has also been suggested to be induced by a chemokine CXCL12 through a calcium-independent pathway [10]. In these studies, contraction was evaluated by observing the shrinkage of collagen lattices on which isolated HSCs were cultured. Since isolated qHSCs are well known to be easily activated by culture on plastic plates [14], the preparations used in these studies might have contained aHSCs when the measurement was performed. In addition, since the shrinkage of collagen lattices develops very slowly, the responses to reagents were measured more than 24 h after their application. It is thus difficult to properly evaluate contraction of qHSCs without denying the involvement of aHSCs with this method.
In the present study, we have developed a novel method to quantitatively evaluate contractile properties in single isolated non-activated HSCs in real-time. The results shown here provide direct evidence for ET-1-induced contraction of non-activated HSCs.

Isolation of HSCs from mice
Male ddY mice (8-12 weeks old; Nihon SLC, Hamamatsu, Japan) were housed under a 12 h light-dark cycle, with food and water available ad libitum and treated as approved by the Institutional Animal Care and Use Committee of the University of Shizuoka and in accordance with the Guidelines for Animal Experiments established by the Japanese Pharmacological Society. HSCs were isolated from mice anesthetized with isoflurane via digestion with pronase (Merck-Millipore, Tokyo, Japan), DNase I (Roche, Basel, Switzerland), and collagenase (Yakult, Tokyo, Japan), followed by density gradient centrifugation with 13% Histodenz (Sigma Aldrich, St. Louis, MO, USA) as described previously [11]. One-day and 7-days primary cultured HSCs were used as non-activated and activated cells, respectively. Cells immunostained with antibody against α-SMA, a marker of aHSC, was markedly increased when cultured in DMEM containing FBS on plastic plates for 7 days (Fig 1I).

Measurement of contraction
The contraction of isolated HSCs was measured on a solid gel mixture containing 2.0-μm fluorescent latex beads (Sigma Aldrich, #L4530-1ML). The gel mixture was made with rat tail type I collagen (Millipore, Darmstadt, Germany) and 2.65× Dulbecco's modified Eagles medium (DMEM, Nissui Pharmaceutical, Tokyo, Japan) and adjusted to pH 7 with 0.5 N NaOH. 150 μL of the gel mixture was plated on a 35 mm glass bottom dish (glass diameter, 14 mm; Matsunami Glass, Osaka, Japan) and flattened. The fluorescent latex beads were diluted 4000-fold with MilliQ water and sonicated for 1 min, and 400 μL of the solution was placed on the gel. The final concentration of collagen in the dish was 1 mg/mL. Collagen was then solidified in a 5% CO 2 incubator at 37˚C. Isolated HSCs, which were cultured in DMEM supplemented with 10% fetal bovine serum (Biowest, Cat No: S1820-500, Lot No: S08048S1820), 100 units/mL penicillin, and 100 mg/mL streptomycin and incubated at 37˚C in a humidified atmosphere with 5% CO 2 for less than 24 h, were seeded on the solid gel mixture 4 h after the gel preparation. For contraction measurement, the dish solution was replaced by HEPES-buffered solution (125 mM NaCl, 5

Measurement of mRNA expression
Total RNA was extracted from non-activated and activated HSCs using the RNeasy Mini kit according to the manufacturer's instructions (Qiagen, Tokyo, Japan). RT-PCR was performed using Qiagen One Step RT-PCR kit (Qiagen, Tokyo, Japan). Primers used for RT-PCR are listed in Table 1.

Statistics
Data are expressed as means ± SEM. Groups were compared using the Holm-Bonferroni multiple comparison test or the unpaired Student's t-test. P values < 0.05 was considered statistically significant.

Measurement of HSC contraction induced by ET-1
Non-activated HSCs were identified by autofluorescence of vitamin A, which was excited at 340 nm and emitted at 550 nm (Fig 1A-1D). One HSC was selected, and the movement of fluorescent latex beads, which were excited at 470 nm and emitted at 505 nm, around this cell was measured as an index of HSC contraction (Fig 1E). Fluorescence images were recorded at a frequency of 0.2 Hz using the AQUACOSMOS system (Hamamatsu Photonics, Hamamatsu, Japan). The recorded images were loaded onto ImageJ software and 8-16 fluorescent beads, which were located between 35 and 150 μm from the cell, were selected for the contraction measurement. The x, y-coordinates of each bead in the continuous images were measured using Particle Track and Analysis plugins for ImageJ and outputted in a text file (Fig 1F). The length of bead movement was calculated from the changes in the coordinates (Fig 1F). The data were averaged and used as the index of cell contraction. Since beads located within approximately 250 μm from the cell moved associated with cell contraction induced by 10 nM ET-1 vide infra, the cell which was more than 350 μm apart from other cells was selected for the measurement to avoid interference between cells (Fig 1G). There was no difference in the contraction between two cells, which were apart more than 350 μm, in the same viewing field ( Fig 1G). We also confirmed that the beads more than 400 μm apart from the cell did not move when the cell contracted ( S2 Fig).
Using the single-cell contraction analysis, we directly demonstrated that ET-1 (0.1, 1, and 10 nM) induced a sustained contraction of isolated mouse non-activated HSCs in a concentration-dependent manner (Fig 2A and S1 Video). The contraction developed slowly, reaching a maximum contraction at around 15 min. Similarly, ET-1 (10 nM) induced a sustained contraction of mouse aHSCs, which were obtained by cultivating for 7 days (Fig 1H). The culturing of isolated primary HSCs in DMEM containing FBS for 5 or 7 days on plastic plates was shown to lead to marked changes in cell shape and an increase in the area of α-SMA-positive staining, which is a marker of aHSC (Fig 1I). We further confirmed that isolated primary HSCs cultured in DMEM containing FBS for 24 h on the collagen gel were kept in a non-activated state, which was demonstrated by immunocytochemistry for cytoglobin, PPARγ, and GFAP, which are known as markers of qHSC (Fig 1J).
RT-PCR analysis showed the mRNA expression of myosin light chain kinase (MLCK) the myosin phosphatase targeting subunit (MYPT1), the regulatory subunit of myosin light chain phosphatase, and non-muscle myosin IIa (NMM IIa) and IIb (NMM IIb) in mouse non-activated HSCs (Fig 2B and 2C). We also confirmed that the contraction of non-activated HSCs induced by ET-1 (10 nM) was suppressed by blebbistatin (10 μM), an inhibitor of NMM II and striated muscle myosin II (Fig 2D and 2E), and by W-7 (1 μM), a calmodulin (CaM) inhibitor (Fig 2F and 2G).

Mechanism of non-activated HSC contraction induced by ET-1
The mRNA expression of both ET A and ET B receptors in non-activated HSCs was confirmed using RT-PCR (Fig 3A). Pharmacological analysis with ET receptor antagonists with different specificities was thus performed to determined which subtypes were involved in the observed contraction. Bosentan (1 μM), a non-selective ET-1 receptor antagonist, abolished the contraction induced by ET-1 (1 and 10 nM; Fig 3B). Ambrisentan (1 μM), a selective ET A receptor antagonist, also significantly reduced the contraction induced by ET-1 (1 nM; Fig 3C, 3D and 3F). In contrast, BQ-788 (100 μM), a selective ET B receptor antagonist, showed no significant effect on the contraction induced by ET-1 (1 nM; Fig 3C and 3E, 3F).

Discussion
Conventionally, the contraction of HSCs has been measured via shrinkage of collagen lattices containing isolated HSCs [9,10,13]. However, isolated qHSCs are well known to be easily activated by culture on plastic plates [14]; therefore, the involvement of aHSCs in the response is hard to be ruled out when using this method. In addition, this method to measure the shrinkage of collagen lattices was not suitable for measuring immediate contraction. In the present study, we have developed a novel method to measure the contraction of single primary nonactivated HSCs cultured on collagen gel. Our results in immunostaining for qHSC markers, i.e., cytoglobin, PPARγ, and GFAP, suggest that HSCs seemed to be kept in a non-activated state by culturing on collagen gel used in the present study at least within 24 h after isolation. Due to the short culture time and the conditions of culture on collagen gel, the activation of HSCs might not be high. Since the activation of HSCs has been shown to revert by culture on soft gel [14,15], the culture on the soft collagen gel, which was used in order to be able to monitor changes in cell length, might also have advantage of keeping HSCs in a non-activated state. Even if there exists aHSC in the preparation, the current method would enable to analyze contractility of single non-activated HSC without any interference by selecting the cell being more than 350 μm apart from other cells. In addition, the current method allows to measure contraction and relaxation in real-time and in immediate response to reagent treatment. We directly demonstrated that ET-1 induces a sustained contraction of non-activated HSCs, which is suggested to be mediated via ET A receptors.
The present study showed the mRNA expression of NMM IIa and IIb in mouse non-activated HSCs and the inhibition of ET-1-induced non-activated HSC contraction by the myosin II inhibitor blebbistatin. These results suggested that non-activated HSC contraction induced by ET-1 is mediated by NMM II. This is the first report suggesting the involvement of NMM II in the contraction of non-activated HSCs. Although aHSCs are known to have contractile activity, NMM II has been suggested not to be involved in ET-1-induced aHSC contraction, as no apparent changes in ET-1-induced contraction of aHSCs were observed in rat aHSCs after knockdown of NMM IIa or IIb [16]. However, in that study, aHSC contraction was evaluated by changes in aHSC-seeded collagen lattice, which was measured 24 h after treatment with ET-1. The method developed in the present study would enable one to measure real-time changes in contraction even in aHSCs and provide more direct evidence for the mechanism of aHSC contraction.
ET-1-induced non-activated HSC contraction was largely dependent on extracellular Ca 2+ . Although the expression of P/Q-type VDCCs in non-activated HSCs was confirmed, the contribution of VDCCs to contraction was determined to be unlikely, since amlodipine, which blocks not only L-type but also P/Q-type VDCC, showed no inhibitory effect on ET-1-induced contraction. Other mediators of Ca 2+ influx include cation channels, such as SOCs. The length of bead movement after the treatment with ET-1 (10 nM) in the absence and presence of blebbistatin (10 μM; D) or W-7 (1 μM; F). Traces in (A), (D), and (F) were obtained from different cells. Bar graphs indicate the maximal contraction within 15 min after treatment with ET-1 (10 nM) in the absence and presence of blebbistatin (10 μM; E) or W-7 (1 μM; G). Each bar represents the mean ± S.E.M. (n = 4). �� P < 0.01, � P < 0.05. https://doi.org/10.1371/journal.pone.0255656.g002

PLOS ONE
Single-cell measurement of ET-1-induced contraction of non-activated HSCs involvement of SOCs in ET-1-induced contraction was also determined to be unlikely, as SOC blockers SKF-96365 [17] and YM-58483 [18] had little effect on this contraction. In contrast, Gd 3+ , which blocks not only SOC at low concentrations, but also non-capacitative Ca 2+ channels at high concentrations, largely reduced ET-1-induced contraction, indicating that cation channels other than SOCs are likely involved in the response. The involvement of cation channels was also supported by the results with amiloride, a broad-spectrum cation channel blocker.
In extracellular Ca 2+ -free conditions, ET-1 still induced contraction of non-activated HSCs. Ca 2+ -independent contraction induced by ET-1 has also been reported in rat aHSCs [19]. Since ET-1induced contraction in Ca 2+ -free solution, especially tonic contraction, was suppressed by the Rho-kinase inhibitor H-1152, we hypothesized the involvement of Rho kinase in Ca 2+ -independent tonic contraction. In rat HSCs, ET-1-induced contraction has been shown to be inhibited by Y-27632, a Rho-kinase inhibitor [20]. In this study, the formation of F-actin stress fibers disappeared in HSCs treated with Y-27632, suggesting that the inhibition of Rho kinase attenuates actin/myosin interaction in HSCs, resulting in impaired contraction. In the present study, however, no morphological changes in non-activated HSCs were observed, perhaps due to the short duration of treatment with H-1152. Alternatively, H-1152 might inhibit a Ca 2+ -sensitization mechanism. Rho kinase mediates Ca 2+ sensitization in smooth muscle cells; it phosphorylates MYPT, a regulatory subunit of myosin phosphatase, resulting in the inhibition of myosin phosphatase [21][22][23]. Ca 2+ sensitization is well known to be involved in ET-1-induced contraction of smooth muscle [24]. We confirmed the expression of MYPT mRNA in non-activated HSCs. Further experiments will be required to determine whether a Ca 2+ -sensitization mechanism mediated by the Rho/Rho kinase pathway is involved in ET-1-induced contraction of non-activated HSCs.
Endothelin receptors are classified into two subtypes, i.e., ET A and ET B receptors. It has been reported that in normal rats, mRNA expression of ET B receptors is detected not only in sinusoidal endothelial cells, Kupffer cells, and hepatocytes, but also in qHSCs. mRNA expression of ET A receptors, however, is detected in qHSC to a larger degree than in these other cell types [25]. The present pharmacological analysis using ET receptor antagonists with different specificity suggested that ET-1-induced contraction of non-activated HSCs is mediated primarily via ET A receptors. However, unlike the non-selective ET receptor antagonist bosentan, the specific ET A receptor antagonist ambrisentan did not completely inhibit ET-1-induced non-activated HSC contraction. Thus, the possibility that ET-1-induced non-activated HSC contraction is partly mediated via ET B receptors could not be ruled out, although the selective ET B receptor antagonist BQ-788 had no effect on the contraction. ET A and ET B receptor cross-talk has been shown to occur in ET-1-induced vascular contraction, which might explain the limited efficacy of selective ET A or ET B receptor antagonists compared to non-selective ET receptor antagonists [26,27].
Under physiologically relevant conditions, liver ET-1 levels may be as low as 0.05 pg/mg tissue [28]. The plasma ET-1 level in patients with liver disease has been shown to be higher than that in healthy subjects; therefore, the role of ET-1 in the liver has been discussed in contribution to the development and progression of liver diseases such as portal hypertension [28]. However, since ET-1 is mainly synthesized in and released from sinusoidal endothelial cells amlodipine (C; 1 μM, an L-and P/Q-type Ca 2+ channel blocker), amiloride (E; 100 μM, a cation channel blocker), or Gd 3+ (G; 1 mM, a cation channel blocker). Traces in (A), (C), (E), and (G) were obtained from different cells. Bar graphs indicate the maximal contraction within 15 min after treatment with ET-1 (10 nM) in control and Ca 2+ -free solution containing 1.5-mM EGTA (B), or in the absence and presence of amlodipine (D; 1 μM), amiloride (F; 100 μM), or Gd 3+ (H; 1 mM). Each bar represents the mean ± S.E.M. (n = 4-5). �� P < 0.01, � P < 0.05. https://doi.org/10.1371/journal.pone.0255656.g004

PLOS ONE
Single-cell measurement of ET-1-induced contraction of non-activated HSCs that neighbor HSCs in the liver, it may affect HSCs through paracrine effects [29,30]. Thus, the local concentration of ET-1 surrounding qHSCs could reach high enough to induce qHSC contraction even under physiological conditions where the plasma ET-1 level is low. We propose a pivotal role for qHSCs in the regulation of hepatic sinusoidal blood flow via their contractile activity under physiological conditions, potentially mediated by ET-1.