Nanchangmycin regulates FYN, PTK2, and MAPK1/3 to control the fibrotic activity of human hepatic stellate cells

Abstract
Editor's evaluation
Introduction
Results
Discussion
Materials and methods
Data availability
References
Article and author information
Metrics

Abstract

Chronic liver injury causes fibrosis, characterized by the formation of scar tissue resulting from excessive accumulation of extracellular matrix (ECM) proteins. Hepatic stellate cell (HSC) myofibroblasts are the primary cell type responsible for liver fibrosis, yet there are currently no therapies directed at inhibiting the activity of HSC myofibroblasts. To search for potential anti-fibrotic compounds, we performed a high-throughput compound screen in primary human HSC myofibroblasts and identified 19 small molecules that induce HSC inactivation, including the polyether ionophore nanchangmycin (NCMC). NCMC induces lipid re-accumulation while reducing collagen expression, deposition of collagen in the extracellular matrix, cell proliferation, and migration. We find that NCMC increases cytosolic Ca²⁺ and reduces the phosphorylated protein levels of FYN, PTK2 (FAK), MAPK1/3 (ERK2/1), HSPB1 (HSP27), and STAT5B. Further, depletion of each of these kinases suppress COL1A1 expression. These studies reveal a signaling network triggered by NCMC to inactivate HSC myofibroblasts and reduce expression of proteins that compose the fibrotic scar. Identification of the antifibrotic effects of NCMC and the elucidation of pathways by which NCMC inhibits fibrosis provide new tools and therapeutic targets that could potentially be utilized to combat the development and progression of liver fibrosis.

Editor's evaluation

The manuscript by Li et al., identifies the polyether ionophore nanchangmycin as a novel anti-fibrotic compound through a comprehensive chemical library screen. Given the lack of clinically available treatments for liver fibrosis, the anti-activation properties of nanchangmycin could represent a novel therapeutic avenue for the treatment of this disease. These studies pave way for future studies evaluating clinical efficacy of nanchangmycin and modified nanchangmycin compounds in the future.

https://doi.org/10.7554/eLife.74513.sa0

Introduction

Chronic liver disease and cirrhosis are the 11th leading cause of mortality in the United States, accounting for over 40,000 deaths annually (Murphy et al., 2021). Liver injuries, including those caused by viral infection, excessive alcohol intake, and nonalcoholic steatohepatitis, can lead to fibrosis, the accumulation of abnormal scar tissue, in the liver (Bataller and Brenner, 2005). If left unchecked, liver fibrosis can progress to cirrhosis and end-stage liver disease (Bataller and Brenner, 2005). HSC myofibroblasts are the primary cell type responsible for liver fibrosis (Friedman et al., 1985; Mederacke et al., 2013). HSCs reside in the perisinusoidal space and represent 5–8% of total cells in the liver. In their quiescent, nonfibrotic state, they store vitamin A as retinol ester in lipid droplets (Geerts, 2001). In response to chronic liver injury, HSCs are activated and trans-differentiate into HSC myofibroblasts, characterized by the loss of lipid droplets, increased contractility, and secretion of ECM proteins, leading to fibrosis (Bataller and Brenner, 2005; Friedman, 2008).

Resolution of liver fibrosis has been observed when the source of liver injury is removed, such as in patients with successful antiviral therapy against hepatitis B or C (Benyon and Iredale, 2000; Bonis et al., 2001; Falize et al., 2006). Two mechanisms can contribute to the reduction of activated HSC myofibroblasts during resolution of liver fibrosis – apoptosis of activated HSC myofibroblasts and reversion of HSC myofibroblasts to a more quiescent phenotype (Friedman, 2008). With regression of fibrosis, 40–50% of HSC myofibroblasts revert to an inactive state in vivo, which is associated with reduced collagen expression (Kisseleva et al., 2012; Troeger et al., 2012). These encouraging observations suggest that liver fibrosis is reversible and targeting HSC myofibroblasts to induce an inactive phenotype may serve as a therapeutic approach to treat patients with liver fibrosis.

Despite many efforts to understand HSC plasticity and target HSC myofibroblasts, there are currently no FDA-approved therapies directed at inhibiting the activity of HSC myofibroblasts. In our previous studies, we developed a small molecule screen to identify compounds that promote HSC inactivation (Chen et al., 2017). In a pilot screen, this approach revealed the antifibrotic effects of tricyclic antidepressants (TCAs). In mechanistic studies, we identified that TCAs inhibit the enzyme acid ceramidase (aCDase). In subsequent studies, we demonstrated that inhibiting aCDase regulates YAP/TAZ-mediated HSC inactivation and reduces fibrogenesis in mouse models and in human precision cut liver slices (Alsamman et al., 2020).

Here, we expanded our screen approximately 10-fold to include 15,867 experimental wells and developed a secondary screen to evaluate primary hits. We find that nanchangmycin (NCMC), a polyether ionophore, promotes HSC inactivation. Furthermore, we demonstrate that NCMC decreases proliferation, migration, and assembly of collagen fibers in the extracellular matrix. In additional mechanistic studies, we show that multiple kinases and signaling pathways are involved in mediating the impact of NCMC on HSC activities, including the FYN, PTK2 (FAK) and MAPK1/3 (ERK2/1) pathways. Taken together, this study defines NCMC as a potent antifibrotic compound that inactivates HSC myofibroblasts and highlights the FYN, PTK2, and MAPK1/3 pathways as potential downstream targets to inhibit liver fibrosis.

Results

A high-throughput small molecule screen identifies compounds that inactivate human hepatic stellate cell myofibroblasts

To identify small molecules that induce reversion of HSC myofibroblasts to an inactive phenotype, we screened 24 compound libraries consisting of 15,867 experimental wells using a high-throughput method to quantify lipid droplet accumulation as an indicator of HSC inactivation (Chen et al., 2017; Figure 1A, Supplementary files 1and and 2). Activated HSCs were seeded in 384-well plates, treated with compounds for 48 hr, fixed, and stained with Bodipy, a fluorescent lipid dye, to analyze the accumulation of lipid droplets. This approach allowed us to screen based on a feature characteristic of quiescent-like inactivated HSCs instead of limiting the readout to expression of a specific gene for the primary screen. Expression of COL1A1 and ACTA2 were then added as a secondary screen to focus on compounds that induced lipid accumulation and reduction of genes that mark the activation and fibrotic activity of HSCs. DMSO and nortriptyline were included as negative and positive controls, respectively on each plate. A scaled value was calculated for each experimental well based on the average percentage of Bodipy-positive cells, toxicity, and reproducibility and was normalized according to negative and positive controls on the same plate to minimize plate-specific effects.

Figure 1 with 2 supplements see all

Download asset Open asset

High-throughput small molecule screen in primary human HSCs.

(A) Overview of the small molecule screen. The number of candidate compounds (# of hits) is indicated at each step. The number in parenthesis represents the number of compounds tested in the final dose response curve (DRC) analysis due to availability. (B) Results of the primary lipid accumulation screen. Each dot indicates the mean scaled value of two replicates for each condition at 48 hr. Red dots represent negative control wells (DMSO), green dots represent positive control wells (nortriptyline, 27 µM), and blue dots represent experimental wells. (C) Results of the secondary mRNA screen. Each dot indicates the mean fold change of *ACTA2* and *COL1A1* after treatment with compounds (normalized to DMSO controls). *PSMB2* was used as the reference gene (n=4). Green dots represent positive hits (<0.5), and blue dots represent non-hits (negative). (D) Dose-response curves were plotted for 39 compounds and were scored by three researchers independently. The mean score for each compound was plotted. The dotted line indicates the score of the positive control nortriptyline. Green dots represent positive hits, and blue dots represent non-hits (negative). This figure has two supplements.

Experimental wells with a scaled value higher than 0.85, a cutoff set as the top 10th percentile of nortriptyline-treated wells on the same plate, were defined as hits. To avoid losing potential hits due to plate-specific effects, the top three experimental wells with the highest scaled values on each plate were also included as hits even if their scaled values did not meet the 0.85 cutoff. A total of 711 experimental wells, containing 464 different compounds, met these criteria and were identified as primary hits (Figure 1B and Supplementary file 3).

To further narrow the candidate list for secondary screening, the 711 hits were separated into 102 clusters based on their chemical structure, with an average cluster size of seven compounds. The highest scoring hit with the most common structure within each cluster was selected as the representative compound for the cluster. Among the 102 representative compounds, ten compounds were removed because they contained pan assay interference structures (Baell and Nissink, 2018) or were themselves frequently identified as hits in screens. One compound was removed because it was a TCA, and we have previously demonstrated that TCAs target the sphingolipid pathway to inhibit HSC activity (Chen et al., 2017). One additional compound was removed because it had the same molecular formula as another selected hit (Supplementary file 4). In addition to representative compounds selected from each cluster, 50 compounds with high scaled values or promising structures were also selected. In total, 140 unique compounds were included in the secondary screen (Figure 1A and Supplementary file 5).

Development of a secondary screening assay in primary human HSCs

Activated HSCs are characterized by increased expression of α-SMA (encoded by ACTA2) and type 1 collagen (encoded by COL1A1) (Bataller and Brenner, 2005; Friedman, 2008). We developed a high-throughput secondary assay to quantify ACTA2 and COL1A1 mRNA levels as indicators of HSC activity. HSCs were treated with compounds for 48 hr and then lysed for multiplexed qRT-PCR to quantify the house-keeping reference mRNA in the same well as ACTA2 or COL1A1. Since a proper reference mRNA is critical for qRT-PCR based assays, we evaluated 18 housekeeping genes, consisting of seven commonly used genes as well as eleven reference genes identified from the literature (Eisenberg and Levanon, 2013). We first analyzed RNA sequencing data from HSCs under multiple conditions (Chen et al., 2017). Among these candidates, GUSB, POLR2A, EMC7, VCP, PSMB2, and VPS29 showed the lowest standard deviation (0.15 or less). Further comparison of expression of these genes in inactivated HSCs (induced by the addition of nortriptyline or ceramide Chen et al., 2017) and culture-activated HSCs revealed that GUSB, POLR2A, EMC7, and PSMB2 had the least fold change in expression (10% or less upon HSC inactivation). Thus, we selected these four reference mRNAs for further evaluation. GAPDH, which is used routinely as a reference control, was also included for comparison (Figure 1—figure supplement 1A). We quantified expression using qRT-PCR in HSC cDNA samples reverse-transcribed from equal amounts of total RNA. PSMB2, which encodes proteasome 20 S subunit beta 2, showed the least variation as indicated by standard deviation and was chosen as the reference mRNA for the secondary qRT-PCR-based screen (Figure 1—figure supplement 1B).

qRT-PCR was performed to quantify ACTA2, COL1A1, and PSMB2 mRNA levels in each sample. Relative fold changes were calculated compared to DMSO control. We defined the following as criteria for compound advancement: 1. Fold change of COL1A1 was reduced to less than 0.5 of DMSO control (FDR <0.05); 2. Fold change of ACTA2 was reduced to less than 0.5 of DMSO (FDR <0.05); 3. Averaged PSMB2 expression was between 0.2 and 2.0 of DMSO (Figure 1C and Supplementary file 5). This last criterion was added to avoid selecting compounds where large changes in PSMB2 expression made it difficult to interpret changes in ACTA2 and COL1A1 expression. Of the 140 compounds, a total of 44 compounds met all three criteria. Five compounds were not commercially available, and 39 compounds were advanced for further analysis.

Next, we evaluated dose response curves (DRCs) for each compound at eight different concentrations, from 10 pM to 10 µM, using a Bodipy lipid accumulation assay similar to that employed in the primary screen. Dose response curves were scored blindly by three researchers (Supplementary file 6 and Materials and methods), and nortriptyline served as a reference. Of the 39 compounds, 19 received an average score that was the same as or higher than nortriptyline controls (Figure 1D, Figure 1—figure supplement 2, and Table 1).

Table 1

Candidates from small molecule screening.

Compound Name	DRC Score	Known function
Cyproheptadine *	1.0	Serotonin antagonist and antihistamine
Emetine *	1.0	Anti-protozoal, inhibitor of Zika and Ebola viruses
Homoharringtonine *	1.0	Translation elongation inhibitor
Nanchangmycin *	1.0	Polyether ionophore antibiotic, inhibitor of Zika virus
Ouabain *	1.3	Na/K-ATPase inhibitor
Digitoxigenin *	1.7	Na/K-ATPase inhibitor
Digoxin *	1.7	Na/K-ATPase inhibitor
Trichostatin A *	1.7	Histone deacetylase inhibitor
AT7867, HMSL10154-101-1	1.8	Multi-kinase inhibitor
PCI-24781 (Abexinostat) *	2.0	Histone deacetylase inhibitor
Dasatinib	2.0	Multi-kinase inhibitor
ChemBridge 35093961	2.3	IKK inhibitor
Homochlorcyclizine dihydrochloride	2.3	Antihistamine
LY2608204	2.3	Glucokinase activator
Vandetanib	2.3	Multi-kinase inhibitor
BIX01294 (hydrochloride hydrate)	2.7	G9a histone methyltransferase inhibitor
HMSL10378-101	2.7	Predicted to target GSK3B at 1 nM (ChEMBL)
Crizotinib	2.8	Multi-kinase inhibitor
Solifenacin succinate	3.0	Muscarinic receptor antagonist

*

Compounds with an EC50 less than 5 μM.

We then selected compounds for additional validation based on their EC50 and DRC scores to identify those considered the most potent. Of the nine compounds selected, two subgroups of compounds were identified based on similar bioactivity – histone deacetylase inhibitors (HDACIs), including trichostatin A and abexinostat, and Na/K-ATPase inhibitors, including ouabain, digitoxigenin, and digoxin. Histone deacetylases are linked to a variety of fibrotic disorders, including liver fibrosis (Pang and Zhuang, 2010). HDACIs, such as MC1568 and Valproate, have been reported to reduce HSC activation and alleviate liver fibrosis in animal models (Yoon et al., 2019). The presence of HDACIs in our final candidate list supports the validity of our screening approach in identifying potential liver fibrosis inhibitors. Na/K-ATPase activity may play a role in non-alcoholic fatty liver disease (Sodhi et al., 2017), but it is not clear how Na/K-ATPases regulate HSC activity and liver fibrosis. Due to the toxicity and narrow therapeutic dose range of cardiac glycosides, which limit their potential application in treatment of liver fibrosis, we decided not to pursue further evaluation of this group of compounds. Nanchangmycin (NCMC), a natural product of Streptomyces nanchangensis, is a polyether insecticidal antibiotic (Sun et al., 2002) and is one of the most potent hits. Studies of NCMC are limited, but it has been shown to have a broad spectrum of antiviral activity against diverse arboviruses (Rausch et al., 2017) and potentially SARS-CoV-2 infection (Dittmar et al., 2020; Li et al., 2020; Svenningsen et al., 2020). It also suppresses breast cancer stem cell activity and inhibits growth of breast cancer and multiple myeloma cells (Huang et al., 2018; Xu et al., 2020). The cellular targets of NCMC remain largely unknown, and its effect on HSC activation and liver fibrosis has not been investigated. Therefore, we selected this compound to further validate its function in regulating HSC activity and explore its mechanism of action.

Nanchangmycin induces lipid accumulation in both human and mouse primary HSCs

First, we confirmed the effect of NCMC on lipid accumulation in primary HSCs isolated from four human donors of different age, gender and race (donor information provided in Materials and methods). We observed that NCMC treatment significantly increased lipid droplet accumulation in primary HSC lines compared to controls (Figure 2A), consistent with HSC inactivation. Compared to nortriptyline (Chen et al., 2017), NCMC exhibited higher potency in HSCs from all four donors, where 100–300 nM of NCMC exhibited similar effects on lipid droplet accumulation to 10 µM nortriptyline (Figure 2B and Figure 2—figure supplement 1). As we switched to NCMC from a different source with higher purity, we found that the new NCMC stock has a lower EC50 in HSCs from human donors and HSCs from C57BL/6 mice (Figure 2C–D and Figure 2—figure supplement 2). These results show that NCMC induces lipid accumulation in both human and murine HSCs.

Figure 2 with 2 supplements see all

Download asset Open asset

NCMC induces lipid accumulation in HSCs.

(A) Representative microscopic images of HSCs treated with DMSO and NCMC for 48 hr. Cells were stained with Bodipy to identify lipid droplets (green) and Hoechst to define nuclei (blue). Scale bar represents 100 µm. (B) Dose response curves for NCMC (purple) and nortriptyline (blue) in HSCs isolated from four different human donors at 48 hr. Dotted line represents the mean percentage of Bodipy-positive cells in DMSO control wells. Error bars represent mean ± SEM (n=12). One experiment was performed independently for each of four donor lines. Curves were generated by fitting the data to a sigmoidal model. The data from the highest two concentrations of NCMC treatment (11 and 33 µM) were not used for fitting due to higher toxicity at these concentrations, as indicated by cell number (Figure 2—figure supplement 1). (C) Dose response curve for NCMC treatment in murine primary HSCs at 48 hr. Dotted lines represent the averaged percentage of Bodipy-positive cells in DMSO-negative control wells (lower) and nortriptyline-positive control wells (10 µM, upper). Error bars represent mean ± SEM (n=6). Data are representative of three independent experiments. (D) Representative images of murine HSCs treated with DMSO and NCMC for 48 hr and stained with Bodipy to identify lipid droplets (green) and Hoechst to define nuclei (blue). Scale bar represents 100 µm. This figure has two supplements.

NCMC inhibits fibrotic gene expression in HSCs

We next quantified the effect of NCMC on ACTA2 and COL1A1 expression in multiple primary human HSC lines. NCMC treatment reduced both ACTA2 and COL1A1 levels at 100 nM or higher (Figure 3A–B). We also observed that NCMC reduced Acta2 and Col1a1 expression at multiple concentrations in murine HSCs at day 2 (Figure 3C). To investigate how NCMC affects the level of collagen deposited into the ECM, we performed the scar-in-a-jar assay to accelerate the process of ECM deposition with addition of molecular crowding reagents and TGF-β (Chen et al., 2009; Good et al., 2019). NCMC treatment significantly decreased collagen staining intensity and fiber area (Figure 3D–E). In addition to two-dimensional (2D) cell culture models, we also tested NCMC’s effect on COL1A1 expression in spheroids consisting of primary human HSCs and primary rat hepatocytes. Both the basal expression of COL1A1 and TGF-β-induced COL1A1 expression were significantly reduced by NCMC treatment (Figure 3F).

Figure 3 with 4 supplements see all

Download asset Open asset

NCMC inhibits expression of fibrotic genes in HSCs.

(**A-B**) Effect of 48 hr NCMC treatment on *ACTA2* and *COL1A1* in HSCs from human donors 1 (A) and 3 (B). Error bars represent mean ± SEM (n=3). Data are representative of three independent experiments. ns indicates not significant, *** indicates p<0.001, and **** indicates p<0.0001 (one-way ANOVA test). (C) Effect of 48 hr NCMC treatment on *Acta2* and *Col1a1* in primary mouse HSCs. Error bars represent mean ± SEM (n=4). Data are representative of three independent experiments. **** indicates p<0.0001 (one-way ANOVA test). (**D–E**) Effect of 48 hr NCMC treatment (1 µM) on collagen deposition in ECM. (D): representative images. Scale bar represents 100 µm. Collagen protein is indicated in green and nuclei for the same field are indicated in blue. (E): quantified results. Error bars represent mean ± SEM (n=4). Data are representative of three independent experiments. **** indicates p<0.0001 (Student’s t-test). (F) qPCR analysis of *COL1A1* in HSC-hepatocyte spheroids treated with NCMC with and without TGF-β (Tβ). Error bars represent mean ± SEM (n=3). One experiment was performed independently for each donor shown. * indicates p<0.05 (Student’s t-test) and ** indicates p<0.01 (Student’s t-test). Analysis was performed on day 3 (3D). (**G–H**) RNA sequencing analysis of HSCs (donor 1) treated with DMSO or 1 µM NCMC for 48 hr. (G) Representative gene sets from the gene set enrichment analysis (GSEA). NES refers to normalized enrichment score. Nom P refers to Nominal P value. Vertical black lines refer to affected genes in the indicated signatures. Red color indicates positive correlation, and blue color indicates negative correlation. (H): Heatmap showing RNA-seq expression for the canonical HSC gene signature (Zhang et al., 2016). All genes from the signature that are expressed in HSCs (with a minimum of five reads) were shown regardless of their expression patterns. Z-score values are also provided in Supplementary file 8. This figure has three supplements.

HSCs were next treated with NCMC and analyzed by RNA-sequencing analysis, which revealed that NCMC broadly affects genes associated with fibrosis. Among the top gene sets negatively enriched in the NCMC-treated group were ECM-related signatures, including ECM structural constituent and collagen-containing ECM, as well as signatures relevant to migration, including contractile fibers. Of note, genes associated with oligosaccharide lipid intermediate biosynthetic process were positively enriched, possibly contributing to the re-accumulation of lipid droplets (Figure 3G and Supplementary file 7). We compared the RNA sequencing data with a canonical HSC gene signature (Zhang et al., 2016), an HSC-specific signature that is highly and uniquely expressed in HSCs and correlates with the extent of fibrosis (Zhang et al., 2016), and the liver cirrhosis signature from Disgenet database (Piñero et al., 2020). We observed that these signatures were significantly negatively enriched (Figure 3—figure supplement 1 and Supplementary file 8). Visualization of the canonical HSC gene signature (Zhang et al., 2016) also demonstrated that many genes that promote liver fibrosis, including those that encode collagens, TGF-β, and PDGF pathway components, are inhibited by NCMC (Figure 3H and Supplementary file 8). The mRNA level of the classic quiescent or inactivated HSC marker PPARG (PPARγ) is also induced with NCMC treatment (Figure 3—figure supplement 2). Taken together, these data suggest that NCMC triggers a global change in HSC gene expression, resulting in a decrease in profibrotic activity. We also used the Enrichr online pathway analysis tool (Chen et al., 2013; Kuleshov et al., 2016; Xie et al., 2021) to analyze the 1904 genes that are significantly (FDR <0.05) upregulated greater than two fold and the 2,620 genes that are significantly (FDR <0.05) downregulated more than twofold by NCMC (Supplementary file 9). The results from analysis of five databases were compared, including Reactome 2016, MSigDB Hallmark 2020, WikiPathway 2021 Human, Elsevier Pathway Collection, and KEGG 2021 Human (Figure 3—figure supplements 3 and 4 and Supplementary file 10). Among the 30 pathways significantly enriched (FDR <0.05) for NCMC-upregulated genes, the unfolded protein response, sterol regulatory element-binding proteins (SREBP) signaling and cholesterol synthesis were identified through analysis of multiple databases. Among the 270 pathways significantly enriched (FDR <0.05) for NCMC-downregulated genes, a few pathways were identified to be represented more than once, for example, TGF-β signaling, Ca²⁺ response and regulation, interferon response, estrogen response, and focal adhesion-PI3K/Akt signaling pathway. These pathways may contribute to the HSC-inactivating effects of NCMC.

NCMC reduces migration and proliferation of HSCs

In addition to secretion of ECM proteins, activated HSCs demonstrate enhanced migration capabilities (Hernandez-Gea and Friedman, 2011). Thus, we performed transwell migration and scratch wound healing assays to evaluate how NCMC affects HSC migration. HSCs were pre-treated with NCMC for two days before seeding in cell culture inserts with permeable membranes. After 6–24 hr, HSCs that migrated through the membrane were stained and counted. NCMC treatment significantly reduced the number of cells that migrated through the membrane compared to DMSO controls (Figure 4A).

Figure 4 with 2 supplements see all

Download asset Open asset

NCMC inhibits HSC migration and proliferation and increases cytosolic calcium concentration.

(A) Transwell migration assay results of HSCs treated with DMSO or 1 µM NCMC for 48 hr. Top: representative images. Scale bar represents 500 µm. Bottom: quantification of migrated cells (n=3 for each experiment). ** indicates p<0.01, and **** indicates p<0.0001 (Student’s t-test). (B) Wound healing assay results of HSCs treated with DMSO or NCMC. HSCs were seeded in complete medium, and immediately after generating the wound field, DMSO and 1 µM NCMC were added. The closure of the wound field was monitored for up to 30 hr as indicated. Left: representative images. White bars highlight the width of the wound field. Scale bar represents 1000 µm. Right: quantification of wound width. (n=2). Data are representative of three independent experiments. (C) Cell count for HSCs treated with DMSO or NCMC over the indicated time in days. Top: human HSCs from donor 1. Middle: human HSCs from donor 3. Bottom: mouse HSCs. Error bars represent mean ± SEM (n=6) but are too small to be visualized. One experiment was performed independently for each HSC line shown. (D) Flow cytometry analysis of Annexin V and propidium iodide (PI) stained HSCs from human donor 1 (top) and 3 (bottom) treated with DMSO or NCMC for 48 hr. Plots are representative of two independent experiments. (E) Flow cytometry analysis of Ki-67 and PI stained HSCs from human donor 1 treated with DMSO or NCMC for 24 hr. Plots are representative of two independent experiments. (F) Measurement of cytosolic calcium level using fluo-4 NW. HSCs from donor 3 were pre-loaded with fluo-4 NW and fluorescent intensity was read immediately after adding compounds (NCMC: nanchangmycin [red], TG: thapsigargin [blue], INMC: ionomycin [green]). The plot demonstrates results from three independent experiments. Solid dots represent mean, and dotted lines represent SEM (n=3). RFU: Relative fluorescence unit. (G) Measurement of cytosolic calcium level after adding NCMC at indicated concentrations in HSCs from donor 3 (magenta: control (no compound, no DMSO), blue: 10 nM, green: 100 nM, red 1 µM, orange: 10 µM). The plot demonstrates results from one experiment for 10 µM and five independent experiments for the other concentrations. Solid dots represent mean, and dotted lines represent SEM (n=3). This figure has two supplements.

We also examined the wound healing capability of HSCs. HSCs were seeded onto plates containing inserts that block cells from accessing and attaching to a strip at the bottom of the well. After removal of the insert, DMSO or 1 µM NCMC was added, and HSCs filled the gap or ‘wound’ field through migration and proliferation. After 30 hr of treatment, the DMSO-treated control cells closed the gap, whereas the gap remained for NCMC-treated cells (Figure 4B). While inhibition of HSC migration by NCMC is likely the major contributor to the difference observed between NCMC treatment and controls at shorter time points, such as 6 hr for the transwell assay, the difference at longer time points could be attributed to reductions in both migration and proliferation.

We next determined how NCMC affects HSC proliferation. We treated HSCs isolated from two human donors with DMSO control and increasing concentrations of NCMC and counted Hoechst-stained nuclei on five consecutive days. NCMC treatment reduced cell proliferation at concentrations of 25 nM and higher, and this effect on proliferation was first evident on day 2 (Figure 4C). A fraction of HSCs undergo apoptosis with the removal of fibrotic stimuli in vivo (Kisseleva et al., 2012), and we evaluated the contribution of apoptosis to the effect of NCMC on day 2. Based on the analysis of Annexin V and propidium iodide (PI) by flow cytometry, NCMC treatment only showed a small increase in the percentage of apoptotic cells in one of two donor HSC lines at 1 µM and no increase at 100 nM. There was also a small increase in necrotic cells in NCMC treatment compared to controls (Figure 4D). In addition, analysis of the proliferation marker Ki-67 and PI showed that NCMC treatment increased the percentage of G0 quiescent cells (Figure 4E and Figure 4—figure supplement 1). Taken together, these results suggest that HSCs enter a more quiescent-like, non-proliferative state upon NCMC treatment, and apoptosis was only observed in a small fraction of cells.

Modulation of Ca²⁺ signaling may contribute to the effect of NCMC on HSC inactivation

NCMC is a polyether ionophore and has been shown to increase cytosolic Ca²⁺ in a cancer cell line (Huang et al., 2018). We examined the activity of NCMC as a calcium ionophore in HSCs. We loaded HSCs with a fluorescent Ca²⁺ indicator (Fluo-4 NW) and treated cells with NCMC. Ionomycin and thapsigargin were included as positive controls (Jones and Sharpe, 1994; Morgan and Jacob, 1994). At 10 µM, NCMC increased cytosolic Ca²⁺ as did ionomycin and thapsigargin (Figure 4F). We also performed the same analysis with ethylene glycol tetraacetic acid (EGTA) to chelate Ca²⁺ and eliminate any contribution from extracellular Ca²⁺ during the assay. Similar to the effect observed with thapsigargin (Ribeiro et al., 2018), the increase of cytosolic Ca²⁺ in response to NCMC was not sensitive to EGTA (Figure 4—figure supplement 2A), suggesting that the immediate increase of cytosolic Ca²⁺ following NCMC treatment is due to release of calcium from intracellular stores. Analysis of a dose response of NCMC demonstrated an increase in cytosolic Ca²⁺ at concentrations as low as 10 nM (Figure 4G and Figure 4—figure supplement 2B).

NCMC reduces COL1A1 expression in HSCs through the FYN pathway

Calcium signaling regulates mitogen-activated protein kinases and non-receptor tyrosine kinases (Filvaroff et al., 1990; Rusanescu et al., 1995; Xia et al., 1996), and we analyzed a kinase array to define kinase signaling molecules modulated by NCMC. HSCs were treated with DMSO or 1 μM NCMC for 1 and 18 hours (Figure 5A). Among the 45 proteins tested, FYN phosphorylation at Y420 was reduced by approximately 40% at both 1 hr and 18 hr. We selected HSPB1 (HSP27), MAPK1/3 (ERK2/1), STAT5A/B, and PTK2 (FAK) to study further in addition to FYN because (1) they also showed decreased phosphorylation at 18 hr, and (2) genes encoding these products are expressed at a relatively high level in HSCs, as indicated from RNA sequencing data, suggesting that these may also be potential mechanistic targets of NCMC in HSCs.

Figure 5 with 1 supplement see all

Download asset Open asset

Multiple kinases mediate the effect of NCMC on *COL1A1* expression.

(A) Kinase array analysis of HSCs treated with DMSO or 1 µM NCMC for 1 or 18 hr. Blue bars indicate mean fold change (n=2) in phosphorylation at specified sites in NCMC-treated cells compared to DMSO-treated cells at 1 hr (light blue) or 18 hr (dark blue). Red triangles indicate the mean RPKM of each corresponding kinase mRNA based on RNA sequencing of HSCs (Chen et al., 2017). The three dotted lines represent 20% increase/decrease or no change in phosphorylation. Kinases highlighted in red were chosen for further investigation. (B) The expression of each candidate kinase gene was depleted using pooled siRNAs, and after 72 hr, *COL1A1* level was determined by qRT-PCR in HSCs isolated from human donor 1 (top), 3 (middle), or 4 (bottom). A non-targeting siRNA is used as a control (NTC). Error bars represent mean ± SEM (n≥4, as indicated by the number of dots). ns indicates not significant (p>0.05), * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, and **** indicates p<0.0001 (one-way ANOVA test). (C) Knockdown efficiency of each siRNA pool in HSCs from human donor 1. Error bars represent mean ± SEM (n≥4). ns indicates not significant (p>0.05), * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, and **** indicates p<0.0001 (Student’s t-test performed for *FYN*, *HSPB1*, and *PTK2* depletion, and one-way ANOVA test performed for *MAPK1*, *MAPK3*, *STAT5A,* and *STAT5B* depletion). One experiment was performed independently for each HSC line shown. This figure has one supplement.

To further investigate the role of these seven kinases in human HSCs, we depleted each kinase using pooled siRNAs in human HSCs from three donors. We observed a consistent reduction of COL1A1 with depletion of FYN, HSPB1, MAPK1, MAPK3, and STAT5B (Figure 5B–C, Figure 5—figure supplement 1), suggesting that each kinase contributes to regulation of COL1A1 expression.

Among the kinases that consistently reduced COL1A1 expression, the reduction of FYN phosphorylation at Y420 was the most prominent following one hour of NCMC treatment, indicating that FYN may mediate the immediate response to NCMC in HSCs. Therefore, we further investigated FYN activity. Western blot of FYN revealed two bands – the upper band was reduced in HSCs treated with NCMC, while the lower band showed little change (Figure 6—figure supplement 1A). Both bands were reduced with depletion of FYN (Figure 6—figure supplement 2), suggesting that both products are encoded by FYN mRNA. FYN phosphorylation was not directly evaluated because antibodies that uniquely recognize phosphorylated FYN are not available. We also probed with a phospho-Src family antibody, which recognizes phosphorylated FYN and other Src family proteins (Figure 6—figure supplement 1B). Two bands of approximately the same size are observed with FYN antibody and pSrc antibody, suggesting that both bands may represent phosphorylated FYN, while it is the product in the upper band that is affected by NCMC treatment. qPCR analysis of NCMC-treated HSCs showed that FYN mRNA level was not affected (Figure 6—figure supplement 3), further indicating that NCMC regulates FYN through a post transcriptional mechanism.

Both depletion of FYN using two different siRNA duplexes (Figure 6A–B) and treatment with 1-Naphthyl PP1, an inhibitor of v-Src, FYN, and ABL (Figure 6C), significantly reduced COL1A1 mRNA level in HSCs. Collagen deposition in the ECM was also impaired by FYN depletion, as indicated by the reduced collagen intensity and fiber area in the scar-in-a-jar assay (Figure 6D). In addition, ectopic expression of a dominant negative Y213A FYN mutant (Kaspar and Jaiswal, 2011) reduced COL1A1 expression in HSCs but did not further decrease COL1A1 level in NCMC-treated cells, suggesting that FYN inhibition is likely to be downstream of NCMC (Figure 6E). Lastly, expression of dominant negative FYN in HSCs resulted in a decrease in phosphorylated MAPK (Figure 6F), suggesting that FYN may crosstalk with the MAPK pathway to exert its function.

Figure 6 with 4 supplements see all

Download asset Open asset

FYN/MAPK pathway regulates collagen expression.

(**A-B**) *FYN* was depleted in HSCs with two siRNAs (si1 and si2). The expression levels of *COL1A1* (A) and *FYN* (B) were analyzed by qPCR after 72 hr in comparison to a non-targeting control siRNA (NTC). Error bars represent mean ± SEM (n=3). At least one experiment was performed independently for each of three donor lines. ** indicates p<0.01, *** indicates p<0.001, and **** indicates p<0.0001 (one-way ANOVA test). (C) HSCs were treated with NCMC or 1-Naphthyl PP1 (PP1) for 48 hr. *COL1A1* level was analyzed by qPCR. Error bars represent mean ± SEM (n=3 for donor 1, and n=6 for donor 3). Data are representative of three independent experiments for donor 1 and experiment for donor 3. **** indicates p<0.0001 (one-way ANOVA test). (D) Effect of FYN-depletion on collagen deposition in ECM. Top: representative images. Scale bar represents 100 µm. Bottom: quantified results. Error bars represent mean ± SEM (n=6). Data are representative of two independent experiments. ** indicates p<0.01, and **** indicates p<0.0001 (one-way ANOVA test). (E) HSCs transduced with control virus or virus containing the cDNA encoding dominant negative mutant FYN (FYN-DN) were treated with DMSO or 100 nM NCMC for 48 hr. Expression of *COL1A1* was quantified by qPCR. Error bars represent mean ± SEM (n=3). Data are representative of three independent experiments. ns indicates not significant (p>0.05), * indicates p<0.05, ** indicates p<0.01, and **** indicates p<0.0001 (two-way ANOVA test). (F) Phospho-MAPK and phospho-PTK2 levels were determined by western blot in control HSCs and HSCs overexpressing DN-FYN. Left: representative Western blot results. Right: quantified results. Representative of two independent experiments. This figure has four supplements.

MAPK1/3 and PTK2 regulate HSC migration

Next, we asked if FYN regulates other HSC phenotypes that are observed with NCMC treatment. HSCs were transfected with pooled FYN siRNAs, and after 3 days, cells were seeded for transwell migration assay. As PTK2 is known to regulate migration of lung fibroblasts (Zhao et al., 2016) and HSCs (Zhao et al., 2017), we included PTK2 siRNAs as a positive control. While PTK2 depletion consistently suppressed migration of HSCs isolated from three different human donors, FYN depletion only reduced migration in one HSC line (Figure 7A–B). We also observed that dominant negative FYN promotes phosphorylation of PTK2 (Figure 6F), while NCMC reduces phosphorylation of PTK2 (Figure 6—figure supplement 4), suggesting that NCMC controls PTK2 phosphorylation and HSC migration through a pathway that is independent of FYN.

Figure 7

Download asset Open asset

MAPK and PTK2 regulate HSC migration.

(**A-B**) Kinases were depleted in HSCs with pooled siRNAs. NTC indicates non-targeting siRNA control. KD indicates the gene transcript that is knocked down. After 72 hr, cells were seeded for transwell migration assay. Migration was assayed at 24 hr for HSCs from donor 1 and 4, and 6 hr for HSCs from donor 3. Top: quantified results. Error bars represent mean ± SEM (n=3). Results are shown for three donor lines. ns indicates not significant (p>0.05), * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, and **** indicates p<0.0001 (one-way ANOVA test). Bottom: representative images. Scale bar represents 500 µm. At least one experiment was performed independently for each of three donor lines. (C) Schematic summarizing the signaling pathways triggered by NCMC treatment and the effect on HSC activity. Red lines represent collagen, orange spots represent NCMC, and green spots represent lipid droplets. This figure has one supplement.

We further tested how other kinases affected by NCMC (MAPK1/3 and HSPB1) regulate HSC migration. Depletion of MAPK1/3 consistently reduced migration across different donors, whereas the influence of HSPB1 depletion varied across HSCs from different donors (Figure 7C). These data indicate that NCMC regulates HSC migration through multiple downstream signaling pathways likely targeting PTK2 and MAPK1/3 as the primary paths to inhibit migration.

Pharmacokinetic studies of NCMC

We next sought to determine whether NCMC itself could serve as an antifibrotic compound in vivo. Previous studies showed that dosing of 1 mg/kg every other day via intraperitoneal (IP) injection for three weeks resulted in decreased breast cancer tumor size (Huang et al., 2018), and that oral administration of 2 mg/kg or 4 mg/kg NCMC daily for 20 days reduces multiple myeloma subcutaneous xenografts in nude mice without affecting liver functions, as indicated by aspartate aminotransferase (AST), alkaline phosphatase (ALP), and alanine aminotransferase (ALT) levels (Xu et al., 2020). To estimate the dose necessary to achieve NCMC exposure in the liver at which an antifibrotic response might be observed, we initially performed pharmacokinetic (PK) studies following IP injection of NCMC at 1 mg/kg to mice. We found a relatively flat plasma concentration-time profile of NCMC after a single dose of 1 mg/kg. The maximum plasma concentration (C_max) of 108±39.7 nM was reached at 4 hr followed by a decrease to less than 50 nM after 12 hr. Based on the analysis of plasma concentration over time, we estimated a terminal elimination half-life (t_1/2) of 5.7±0.87 hr. As anticipated from the flat concentration-time curve, the calculated volume of distribution (Vss/F) is nearly 10 L/kg, indicating substantial distribution into tissue.

We next repeated IP dosing daily at 1 mg/kg for 4 days to determine how this affects plasma and tissue concentrations. The plasma concentration-time profile measured after the fourth dose was comparable to that after single dose, giving no hint of an accumulation of the compound after multiple dosing (Table 2). We detected 10-fold higher NCMC concentration in fat tissue compared to plasma, whereas it was undetectable in liver and muscle. The trough total plasma concentration was in the range of EC50 in cell culture, which would result in a lower effective concentration when taking into consideration the fraction of NCMC bound to serum proteins, which we measured at >99% (the reported plasma concentrations refer to the total = free + bound plasma concentrations). Further, hepatic microsomes from mice showed evidence of degradation of NCMC by phase I liver metabolism, as indicated by the short half-life (37.8 min) and high derived intrinsic clearance (36.7 µl/min/mg protein).

Table 2

Pharmacokinetics of NCMC in vivo.

Plasma and tissue concentrations of NCMC were measured at the indicated time points after once daily intraperitoneal injection of 1.0 mg/kg NCMC to male C57/B6N mice (n=4) for 4 days.

Day	Time[h]	Tissue	N	mean[nmol/L]	SD[nmol/L]	CV[%]
1	4	Plasma	4	109.7	23.19	21.1
1	24	Plasma	4	38.25	3.18	8.3
2	24	Plasma	4	32.15	3.99	12.4
4	2	Plasma	4	69.43	2.92	4.2
4	4	Plasma	4	66.70	13.77	20.6
4	8	Plasma	4	80.13	13.55	16.9
4	12	Plasma	4	67.58	12.54	18.6
4	24	Plasma	4	35.25	8.97	25.4
4	24	Muscle	4	NOP	na	na
4	24	Liver	4	NOP	na	na
4	24	Fat	4	375.3	145.2	38.7

N: number of mice, SD: standard deviation, CV: coefficient of variance, NOP: no peak, na: not applicable

In summary, the pharmacokinetic data suggest that NCMC will not reach sufficient exposure in the liver to achieve hepatic antifibrotic effects in vivo.

Discussion

Liver fibrosis is the major driver of liver failure in all etiologies of chronic liver disease, and the degree of fibrosis is the strongest predictive factor for liver-related mortality (Anstee et al., 2019). Current therapies for liver fibrosis focus on eliminating the underlying etiology. However, there is a lack of effective treatment for several chronic liver diseases, such as non-alcoholic fatty liver disease, which affects one in four people worldwide (Younossi et al., 2016), and primary sclerosing cholangitis (Karlsen et al., 2017). Therefore, there is an urgent need to develop antifibrotic therapies. Activation of HSCs in the setting of chronic liver injury represents a critical event in fibrosis, as activated HSC myofibroblasts are the primary source of collagen production and excessive extracellular matrix deposition (Friedman et al., 1985; Maher and McGuire, 1990; Mederacke et al., 2013). With evidence that the scarring process in the liver is reversible (Sun et al., 2020) and that HSC myofibroblasts demonstrate plasticity and can revert to an inactive state (Kisseleva et al., 2012; Troeger et al., 2012), there is increasing enthusiasm for development of approaches to promote HSC inactivation as a therapeutic strategy to treat liver fibrosis. Although some therapies under investigation in clinical trials, including PPAR agonists and TGF-β inhibitors, are anticipated to promote HSC inactivation, none have yet been recognized as effective antifibrotic agents (Guo and Lu, 2020).

Therefore, with the goal to identify new compounds with antifibrotic potential and novel pharmacological targets for the treatment of liver fibrosis, we performed a small molecule compound screen using primary human HSC myofibroblasts. Combining high-content microscopy imaging and high-throughput qPCR screening approaches as well as filtering methods that take into consideration both the potency and diversity of the candidates’ chemical structures, we screened 15,867 experimental wells and identified 19 top candidates that are potentially more potent than the previously discovered HSC-inactivating compound nortriptyline (Chen et al., 2017). The top 19 compounds were listed in Table 1, and the full datasets of the screening results are provided as Supplementary files, which can serve as a resource for future studies. Further studies to investigate the compounds identified by the screen will deepen our understanding of HSC biology and allow identification of additional genes and pathways that could be targeted to reduce liver fibrosis. There are two potential limitations of this screen. First, although lipid droplet re-accumulation is recognized as a feature of quiescent and inactivated HSCs (Zisser et al., 2021), it is possible that partially inactivated HSCs may not accumulate enough lipid droplets to pass our primary screen cutoff, and therefore compounds that act to reduce collagen expression or other fibrosis-related aspects of HSCs without affecting lipid accumulation may be missed in this screen. Second, the lipid accumulation and mRNA screens were conducted with primary human HSCs isolated from the same donor 1, therefore we performed the DRC screen in primary human HSCs isolated from donor 2, so that the final screen hits are not restricted to a single HSC isolate. We validated NCMC in primary HSC lines from two additional donors, but for the other positive hits in the screen, further validation in HSC from additional donors is needed.

We focused on NCMC because it strongly induced HSC inactivation, and the activity of NCMC was poorly understood. NCMC belongs to a group of naturally occurring polyether ionophores, which consist of over 120 known members (Huang et al., 2018). Among them, the compound monensin shares a similar chemical structure to NCMC and is also a positive hit in our primary screen. It was grouped in the same chemical cluster as NCMC, but NCMC was selected as the representative compound for this cluster in the screen because NCMC had a higher scaled value (Supplementary file 4). This group of compounds demonstrates antibacterial, antifungal, antiparasitic, antimalarial, antiviral, anti-inflammatory activities and cytotoxicity in cancer cells (Kevin Ii et al., 2009). Although some polyether ionophores have been employed as veterinary antibiotics, none have been used as antibiotics in human, possibly due to concerns about toxicity (Huczyński, 2012). Indeed, we observe that NCMC demonstrated substantial cytotoxicity in cell culture at concentrations higher than 10 µM (Figure 2—figure supplement 1). However, given that the EC50 of NCMC in lipid accumulation assay is in the range of 10–300 nM without inducing apoptosis, NCMC or compounds with similar structure may have potential as antifibrotic therapy within optimized therapeutic doses.

Despite the extensive study of some polyether ionophores, limited data are available describing the activity and mechanism of action of NCMC in mammalian cells. A screen for bioactive inhibitors of the Otub1/c-Maf axis in multiple myeloma cells demonstrated that NCMC induces c-Maf polyubiquitination and proteasomal degradation in the presence of Otub1 (Xu et al., 2020). A-130-A, a close analog of NCMC, also inhibited the Wnt/β-catenin pathway and induced autophagy, and both A-130-A and NCMC increased cytosolic Ca²⁺ and reactive oxygen species (ROS) as well as enhanced the permeability of the mitochondrial inner membrane to H⁺ and K⁺ (Huang et al., 2018). To examine a wide variety of signaling pathways that may be affected downstream of NCMC, we performed a phospho-kinase array analysis to measure the phosphorylation of 37 kinases at a total of 43 different sites and the total protein expression of β-catenin and HSP60. Seven kinases, including FYN, HSPB1, MAPK1, MAPK3, STAT5A, STAT5B, and PTK2, were selected and further tested because their phosphorylated protein was reduced by NCMC treatment, and all were relatively abundant in HSCs. When depleted individually in HSCs from multiple donors, most of these kinases reduced COL1A1 level consistently (Figure 5B), suggesting that each may play a role in mediating NCMC’s effect on collagen expression. It is unclear why STAT5A siRNAs, which effectively depleted STAT5A but also reduced STAT5B mRNA level, did not significantly affect COL1A1 expression, considering that reduction of STAT5B alone to a similar level by STAT5B siRNAs did demonstrate an inhibitory effect (Figure 5B–C and Figure 5—figure supplement 1). It is possible that the expression of STAT5B protein is inhibited more efficiently in HSCs transfected with STAT5B siRNAs than those transfected with STAT5A siRNAs despite the similar mRNA levels. STAT5A/B homo- and heterodimers could have different individual DNA-binding specificities (Maurer et al., 2019), and the ratio of homo- to heterodimers could also affect transcription of the COL1A1 gene.

Our investigations of FYN revealed that depletion or inhibition of FYN activity suppresses collagen expression in primary human HSCs and deposition of collagen in the ECM, but regulation via FYN did not explain all the effects observed with NCMC. The regulation of collagen expression by FYN is in agreement with a recent study demonstrating that FYN depletion and inhibition in the presence of TGF-β reduces collagen I expression in immortalized human and rat HSC lines (Du et al., 2020). This study also observed that FYN depletion and inhibition in the presence of TGF-β reduced HSC migration (Du et al., 2020), however, our data did not show a consistent effect on migration with depletion of FYN. As cell migration is controlled by a complex signaling network, the difference in the basal activity of the signaling pathways up- or down-stream of FYN may account for the observed differences. In contrast, PTK2 and MAPK1/3 depletion showed a more robust and consistent inhibitory effect among all HSC lines tested, suggesting that these kinases may serve as the critical nodes regulating HSC migration. Depletion or inhibition of FYN or MAPK suppresses HSC proliferation (Du et al., 2020; Pagès et al., 1993), and PTK2 regulates proliferation in many cell types (Zhou et al., 2019). Thus, it is likely that modulation of multiple kinases contributes to the anti-fibrotic effect of NCMC, although the involvement of each kinase may vary depending on the cellular context.

We found that NCMC increases cytosolic Ca²⁺. Although it has not been demonstrated experimentally, it is suspected that NCMC may increase cytosolic Ca²⁺ levels by disrupting Na⁺/Ca²⁺ exchange (Huang et al., 2018). Of note, the Na⁺/K⁺-ATPase inhibitors identified in our screen can also increase cytosolic Ca²⁺ levels (Tian and Xie, 2008), and increased cytosolic Ca²⁺ has been observed to inhibit MAPK1/3 in fibroblasts (Bosch et al., 1998; Chuderland et al., 2020; Cook et al., 1997). As a ubiquitous second messenger with wide-ranging physiological roles, cytosolic Ca²⁺ levels may be a key factor mediating the downstream anti-fibrotic activity of NCMC, but further investigations are needed to unravel the complete signaling cascade.

NCMC has been shown to inhibit tumor growth in a model where breast cancer cells were injected subcutaneously and a multiple myeloma subcutaneous xenograft model (Huang et al., 2018; Xu et al., 2020). Using the same administration method as described by Huang et al. (1 mg/kg NCMC delivered via intra-peritoneal injection) and increasing the dosing frequency to once daily, we were unable to detect NCMC in liver or muscle, while we observed increased concentrations of NCMC in fat compared to plasma. Our in vitro experiments with mouse liver microsomes support metabolization of NCMC in hepatocytes, which may counteract a potential local depot formation in the liver. As NCMC is not detectable in the liver and has a low plasma concentration with high plasma protein binding, we expect HSC exposure to NCMC to be very low. Consequently, chemical modifications of NCMC are needed to come to a drug suitable for the investigation of liver fibrosis in vivo. It is worth screening chemical modifications of NCMC for a more drug-like derivative in the future. Alternatively, molecules similar in structure to NCMC may be evaluated to identify additional compounds with similar activity in HSCs and greater metabolic stability.

Interestingly, salinomycin, another polyether ionophore, has been reported to protect hepatocytes against CCl₄-induced oxidative stress and liver injury in vivo (Kim et al., 2018). Salinomycin was tested in our primary screen and showed a modest effect on lipid accumulation but did not reach the cutoff for selection. A pharmacokinetic study of salinomycin revealed rapid metabolism by liver microsomes and showed that CYP3A4 is the major metabolizing enzyme (Resham et al., 2015). Concomitant administration of ketoconazole, a selective CYP3A4 inhibitor with salinomycin in rats increased the systemic exposure of salinomycin to seven fold and the C_max to three fold (Resham et al., 2015). These findings suggest that in vivo pharmacokinetics of polyether ionophores could also be improved via concomitant inhibition of the metabolizing enzyme.

In summary, this study has identified NCMC as a compound that increases cytosolic Ca²⁺ and regulates multiple kinases, including FYN, PTK2, and MAPK1/3 to drive the inactivation of HSC myofibroblasts (Figure 7C). Targeting an individual component of this complex network may suppress certain cellular activities and contribute to HSC inactivation, but it may be necessary to synergistically manipulate multiple targets to achieve antifibrotic effects among genetic diversity, as observed with responses in multiple primary human HSC lines. By regulating multiple signaling pathways, NCMC confers a more robust impact than observed with inhibition of only one pathway, and with structural modifications to improve delivery to the liver, similar compounds could represent a more effective strategy to reduce the progression of liver fibrosis.

Materials and methods

Key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Gene (Homo sapiens)	FYN (dominant negative mutant with Y213A mutation)	Amplified from the Addgene plasmid pRK5 DN-Fyn	Addgene plasmid pRK5 DN-Fyn (#16033)
Cell line (Homo sapiens)	HEK-293	ATCC	CRL-1573
Biological sample (Homo sapiens)	Primary human hepatic stellate cells (donor 1)	Isolated from human nonparenchymal liver cells (NPCs) purchased from Lonza (cat# HUCNP)	Lonza ID: 4105	Age: 45; Gender: Male; Race: Caucasian
Biological sample (Homo sapiens)	Primary human hepatic stellate cells (donor 2)	Isolated from human nonparenchymal liver cells (NPCs) purchased from Lonza (cat# HUCNP)	Lonza ID: 4270	Age: 35; Gender: Male; Race: Caucasian
Biological sample (Homo sapiens)	Primary human hepatic stellate cells (donor 3)	Lonza (cat# HUCLS)	Lonza ID: 180761	Age: 57; Gender: Female; Race: Caucasian
Biological sample (Homo sapiens)	Primary human hepatic stellate cells (donor 4)	Lonza (cat# HUCLS)	Lonza ID: 182821	Age: 24; Gender: Female; Race: African American
Biological sample (Homo sapiens)	Primary human hepatic stellate cells (donor 5)	Isolated from human nonparenchymal liver cells (NPCs) purchased from Lonza (cat# HUCNP)	Lonza ID: 4258	Age: 51; Gender: Male; Race: African American
Antibody	Anti-collagen type I (Mouse monoclonal)	Sigma	C2456	(1:1000)
Antibody	Anti-FYN (Rabbit polyclonal)	Cell Signaling Technology	#4023	(1:1000)
Antibody	Anti-phospho-Src family (Rabbit monoclonal)	Cell Signaling Technology	#6943	(1:1000)
Antibody	Anti-phospho-p44/42 MAPK (Rabbit monoclonal)	Cell Signaling Technology	#4370	(1:1000)
Antibody	HRP-β-Actin antibody (Mouse monoclonal)	Santa Cruz Biotechnology	sc-47778 HRP	(1:5000)
Antibody	goat anti-rabbit IgG secondary antibody (Goat polyclonal)	Invitrogen	#32460	(1:3000)
Antibody	donkey anti-mouse Alexa Fluor 488 secondary antibody (Donkey polyclonal)	Invitrogen	A-21202	(1:500)
Recombinant DNA reagent	pLJM1-eGFP (plasmid)	Addgene	plasmid# 19319
Recombinant DNA reagent	pRK5 DN-Fyn	Addgene	plasmid# 16033
Recombinant DNA reagent	pLJM1-DN-FYN (plasmid)	This paper		This is a plasmid generated by this study for the expression of dominant negative (Y213A) FYN in HSCs using lentivirus system. Refer to the Materials and Methods section for more information.
Sequence-based reagent	FYN-forward primer	This paper	PCR primer	This is the forward PCR primer for cloning human FYN gene. The sequence is 5’-CAT GCTAGC GCCACC ATGGGCTGTGTGCAATGTAAGG-3’. Refer to the Materials and Methods section for more information.
Sequence-based reagent	FYN-reverse primer	This paper	PCR primer	This is the reverse PCR primer for cloning human FYN gene. The sequence is 5’-AGC GAATTC TTACAGGTTTTCACCAGGTTGGTAC-3’. Refer to the Materials and Methods section for more information.
Sequence-based reagent	siRNA: non-targeting control	Horizon Discovery	D-001210–05	siGENOME
Sequence-based reagent	Pooled siRNAs: FYN	Horizon Discovery	MQ-003140–04	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: HSPB1/HSP27	Horizon Discovery	M-005269–01	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: MAPK3/ERK1	Horizon Discovery	M-003592–03	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: MAPK1/ERK2	Horizon Discovery	M-003555–04	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: PTK2/FAK	Horizon Discovery	M-003164–02	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: STAT5A	Horizon Discovery	M-005169–02	siGENOME SMARTpool
Sequence-based reagent	Pooled siRNAs: STAT5B	Horizon Discovery	M-010539–02	siGENOME SMARTpool
Sequence-based reagent	siRNA#1: FYN	Horizon Discovery	D-003140–10	siGENOME
Sequence-based reagent	siRNA#2: FYN	Horizon Discovery	D-003140–24	siGENOME
Commercial assay or kit	TaqMan Real-time PCR Assay: Human ACTA2	ThermoFisher Scientific	Hs00426835_g1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human COL1A1	ThermoFisher Scientific	Hs00164004_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human FYN	ThermoFisher Scientific	Hs00176628_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human HSPB1/HSP27	ThermoFisher Scientific	Hs00356629_g1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human MAPK1/ERK2	ThermoFisher Scientific	Hs01046830_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human MAPK3/ERK1	ThermoFisher Scientific	Hs00385075_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human PSMB2	ThermoFisher Scientific	Hs01002946_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human PTK2/FAK	ThermoFisher Scientific	Hs01056457_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human STAT5A	ThermoFisher Scientific	Hs00559643_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Human STAT5B	ThermoFisher Scientific	Hs00560026_m1
Commercial assay or kit	TaqMan Real-time PCR Assay: Mouse Acta2	ThermoFisher Scientific	Mm00725412_s1
Commercial assay or kit	TaqMan Real-time PCR Assay: Mouse Col1a1	ThermoFisher Scientific	Mm00801666_g1
Commercial assay or kit	TaqMan Real-time PCR Assay: Mouse Psmb2	ThermoFisher Scientific	Mm00449477_m1
Commercial assay or kit	Proteome Profiler Human Phospho-Kinase Array Kit	R&D Systems	ARY003B
Chemical compound, drug	Nanchangmycin (NCMC)	Selleck Chemicals	S1450	Used for the initial confirmation of dose response curves in multiple HSC lines
Chemical compound, drug	Nanchangmycin (NCMC)	Adooq	A10621	Used for all other follow-up experiments
Chemical compound, drug	1-Naphthyl PP1	Tocris	#3603
Chemical compound, drug	Thapsigargin	Sigma-Aldrich	T9033
Chemical compound, drug	Ionomycin	Biogems	#5608212
Software, algorithm	BIOVIA Pipeline Pilot	Dassault Systèmes
Software, algorithm	GSEA	UC San Diego and Broad Institute	http://www.gsea-msigdb.org/gsea/index.jsp
Software, algorithm	FastQC	Babraham Bioinformatics	v 0.11.8
Software, algorithm	RSEM	https://github.com/deweylab/RSEM; Dewey Lab, 2020	v 1.3.1
Software, algorithm	Enrichr	https://maayanlab.cloud/Enrichr/

Share this article

Cite this article

High-throughput small molecule screen in primary human HSCs.

Candidates from small molecule screening.

NCMC induces lipid accumulation in HSCs.

NCMC inhibits expression of fibrotic genes in HSCs.

NCMC inhibits HSC migration and proliferation and increases cytosolic calcium concentration.

Multiple kinases mediate the effect of NCMC on COL1A1 expression.

FYN/MAPK pathway regulates collagen expression.

MAPK and PTK2 regulate HSC migration.

Pharmacokinetics of NCMC in vivo.

Author details

Wenyang Li

Present address

Contribution

Competing interests

Jennifer Y Chen

Present address

Contribution

Competing interests

Cheng Sun

Contribution

Competing interests

Robert P Sparks

Contribution

Competing interests

Lorena Pantano

Contribution

Competing interests

Raza-Ur Rahman

Contribution

Competing interests

Sean P Moran

Contribution

Competing interests

Joshua V Pondick

Contribution

Competing interests

Rory Kirchner

Contribution

Competing interests

David Wrobel

Contribution

Competing interests

Michael Bieler

Contribution

Competing interests

Achim Sauer

Contribution

Competing interests

Shannan J Ho Sui

Contribution

Competing interests

Julia F Doerner

Contribution

Competing interests

Jörg F Rippmann

Contribution

Competing interests

Alan C Mullen

Contribution

For correspondence

Competing interests

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Categories and tags

Research organism

Further reading