Chaetocin (#C9492, Sigma-Aldrich, St Louis, MO, United States), z-VAD-fmk (#ALX-260–020, Enzo, New York, NY, United States), N-acetyl-L-cysteine (NAC) (#A7250, Sigma-Aldrich) and SP600125 (#S5567, Sigma-Aldrich) were dissolved in DMSO to form the 10mM, 50mM, 0.5M and 20 mM solution respectively. These reagents were stored at −20°C. Primary antibodies against PARP (#9542), caspase-3 (#9662), cleaved-caspase-3 (#9661), caspase-8 (#9746), caspase-9 (#9508), BCL-2 (#15071), BCL-XL (#2764), MCL-1 (#4572), XIAP (#14334), JNK (#9252) and phospho-JNK (Thr183/Tyr185) (#4668) were purchased from Cell Signaling Technology (Beverly, MA, United States). Anti-c-Jun (#ab131497), anti-Ser63 phosphor-c-Jun (#ab28807), anti-CD47 (#ab9089) anti-α-tubulin (#ab233661) and β-actin (#ab8226) were purchased from Abcam (Cambridge, MA, United States), FITC anti-human CD47 (#323106) was purchased from Biolegend (San Diego, CA, United States). Anti-mouse immunoglobulin G (#B900620) and anti-rabbit immunoglobulin G (#B900610) horseradish peroxidase-conjugated secondary antibodies were purchased from Proteintech Group (Chicago, IL, United States).

LIM1215 cell line was obtained from the European Collection of Authenticated Cell Cultures (Salisbury, Wiltshire, United Kingdom). The other CRC cell lines were purchased from Culture Collection of Chinese Academy of Science (Shanghai, China). The 5-FU-resistant HCT-15/5FU-R cell line was established from HCT-15 as previously described (Wen et al., 2015; Wen et al., 2016; Wang et al., 2020). DLD-1, LS174T, Caco-2, LIM1215, HCT-8, SW620, HCT116 and SW480 cells were cultured in RPMI 1640 (Gibco Life Technologies, Carlsbad, CA, United States). HT-29, LoVo, RKO, HCT-15 and HCT-15/5FU-R cells were cultured in DMEM (Gibco Life Technologies) supplemented with 10% fetal bovine serum (Gibco Life Technologies), 100 units/mL penicillin and 10 μg/ml streptomycin (Gibco Life Technologies). Incubations were controlled in a humidified atmospheric environment of 5% CO2 at 37°C.

To test CRC cell viability after chaetocin treatment, CCK-8 assay (Nanjing KeyGen Biotech Co., Ltd., Nanjing, Jiangsu, China) was performed. Cells seeded in 96-well plates were left to grow overnight. The next day, chaetocin at various concentrations were added, and incubated for 24 h. Cells were assessed after 4 h of incubation with 10 μL of CCK-8 solution. The results were carried out using Varioskan Flash multimode reader (Thermo Fisher Scientific, Waltham, MA, United States) at 450 nm.

500 cells/well of HCT116 and LS174T cells and 1,000 cells/well of LIM1215 and HCT–15/5-FU-R cells were seeded in 6-well plates. Chaetocin at various concentration were then added. Plates were left for incubation for different duration. Before imaging, we fixed cells using ice-cold methanol for 5 min and then cells were stained using 0.1% crystal violet. Colonies were counted from images obtained by Epson scanner (Suwa, Nagano, Japan).

To dynamically monitor cell proliferation, xCELLigence system from Roche Applied Science (Mannheim, Baden-Wuerttemberg, Germany) was utilized. Briefly, cells in 100 μL of media were seeded into E-plate and incubated overnight. Chaetocin treatment took place the next day. The impedance measurement was performed according to the manufacturer’s manual. The measurable impedance should increase when cells proliferate.

FACSCanto II flow cytometry (BD Biosciences) with Propidium Iodide (PI, BD Biosciences, Franklin Lakes, NJ, United States) staining was performed. After chaetocin treatment, cells were collected, washed before fixed overnight with 66% cold ethanol at 4°C. After discarding the ethanol, cells were washed with PBS, followed by staining with PI.

Flow cytometry with Annexin V-FITC/PI staining (Nanjing KeyGen Biotech Co., Ltd.) was used to identify the apoptotic population of CRC cells. After chaetocin treated, both suspended and attached cells were collected, washed with PBS and stained in a working solution (500 μL of binding buffer with 5 μL of Annexin V-FITC and 5 μL of PI) for 15 min at room temperature in the dark. FACSCanto II flow cytometry (BD Biosciences) were then carried out, while cells with positive Annexin V-FITC staining were considered as apoptotic cells.

HCT116 cells and LS174T were treated with 0.5 and 1 μM chaetocin for 24 and 12 h respectively. After PBS washed, cells were then incubated with anti-CD47-FITC antibody for 15 min on ice in the dark. The expression of cell surface CD47 was analyzed using FACSCanto II flow cytometry (BD Biosciences).

Total RNA was extracted from chaetocin-treated CRC cells using Trizol reagent (Ambion, Carlsbad, CA, United States). cDNA was then synthesized from 500 ng RNA of each sample using the PrimeScript RT reagent Kit (TaKaRa, Dalian, Liaoning, China). Real-time PCR using SYBR Premix Ex Taq II Kit (TaKaRa) was performed according to manufacturer’s instruction. The specific primers for real-time PCR are as follows: CD47 forward, 5′-TGG TGG GAA ACT ACA CTT GCG-3′; CD47 reverse, 5′- CGT GCG GTT TTT CAG CTC TAT-3′; actin forward, 5′-GTG ACG TTG ACA TCC GTA AAG A-3′; actin reverse, 5′- GCC GGA CTC ATC GTA CTC C-3′; actin was chosen as endogenous control and the CD47 gene expression was analyzed using 2^−ΔΔCt method. The CD47 gene expression is shown as fold change of control.

The mitochondrial membrane potential of CRC cells was measured using the JC-1 mitochondrial membrane potential assay kit (Nanjing KeyGen Biotech Co., Ltd.), following the manufacturer’s protocol. In brief, chaetocin treated cells were stained at 37°C for 20 min, then rinsed, using solutions provided in the kit. The fluorescence intensity results were obtained using FACSCanto II flow cytometry (BD Biosciences), whilst mitochondrial JC-1 aggregates which stained red were considered as mitochondria from health cells.

The cells were incubated with 10 μM of 2′,7′-dichlorofluorescin diacetate (DCFH-DA) (Beyotime Institute of Biotechnology, Shanghai, China), a fluorescent probe, at 37°C for 30 min. Following chaetocin treatments, cells were washed with serum-free media and resuspended in 500 μL of PBS before fluorescence measurements using flow cytometry (BD Biosciences).

RIPA lysis buffer (Cell Signaling Technology) containing protease and phosphatase inhibitors (Nanjing KeyGen Biotech Co., Ltd.) was used to extract proteins from chaetocin treated cells. BCA assay (Pierce, Thermo Fisher Scientific) was implemented to ensure a normalised protein concentration. Equal amounts of protein were resolved by SDS-PAGE and were then transferred to nitrocellulose membranes (Merck Millipore, Burlington, MA, United States). Next, the membranes were first blocked by 5% nonfat dry milk, followed by primary antibody incubation overnight at 4°C. After washing the membranes with TBST the next day, HRP-conjugated secondary antibody was applied and left for incubation for 1 h at room temperature before ECL detection. α-tubulin was used as loading control.

Preparation of human macrophages: monocyte-differentiated macrophages were prepared as described previously (Lou et al., 2019). Monocytes extraction from blood of healthy adult donor was done using MACS using anti-CD14 microbeads (Miltenyi Biotec, Auburn, CA, United States). Monocytes were then cultured in RPMI 1640 (Gibco Life Technologies) supplemented with 10% fetal bovine serum (Gibco Life Technologies), 100 units/mL penicillin and 10 μg/ml streptomycin (Gibco Life Technologies), 200 mM glutamine (Gibco Life Technologies) and 25 mM HEPES (Gibco Life Technologies) and treated with 25 ng/ml human recombinant macrophage colony-stimulating factor (M-CSF) for 7 days to differentiate to macrophages.

Preparation of tumor cells: Live HCT116 cells were stained with 5 μM carboxyfluorescein diacetate succinimidyl ester (CFSE) at 37°C for 15 min after 0.5 μM chaetocin treatment for 24 h. After washed with ice-cold D-PBS, stained cells were resuspended in serum-free RPMI 1640.

For in vitro phagocytosis assay, 1 × 10⁵ macrophages were added to 2 × 10⁵ CFSE-stained tumor cells per well in the 24-well plate for 24 h incubation. Cells were then stained with CD14-APC (Ebiosciences, San Diego, CA, United States) and phagocytosis was analyzed by flow cytometry. CD14-APC plus CFSE positive cells were determined as macrophages which had successfully phagocytized tumor cells.

5 × 10⁶ HCT116 cells were injected subcutaneously into the flank of five-week-old female BALB/c nude mice from Guangdong Laboratory Animals Monitoring Institute. After waited 1 week for measurable tumors to form (approximately 50 mm³), the animals were divided randomly into two groups (n = 6 each group) which was either treated with vehicle (10% DMSO, 20% cremophor EL and 70% NaCl, i. p.) or chaetocin (0.5 mg/kg, i. p.) every day for 18 days. Every other day after treatment began, the body weight of the animals and tumor size were taken. The tumor volumes were estimated as the following formula: a² × b × 0.5, where a is the smallest diameter and b is the diameter perpendicular to a. At the end of the study, the tumors from sacrificed mice were removed, weighed and stored for further experiments. All animal studies conducted were approved by the Institutional Animal Care and Use Committee of Guangdong Laboratory Animals Monitoring Institute.

All experiments were performed at least in triplicate, and the results were shown as mean ± SD where applicable. One-way analysis of variance followed by Tukey’s test by GraphPad Prism 7.00 (GraphPad Software Inc., La Jolla, CA, United States) was used to analyze significant differences among the results. *p < 0.05 was considered statistically significant.

Chaetocin has a thiodioxopiperazine structure as shown in Supplementary Figure 1 (Wen et al., 2019). Firstly, the cytotoxic effects of chaetocin were tested on various CRC cells. After treated with chaetocin for 24 h at increasing concentrations, from 0 to 2 μM, CCK8 results showed a progressive inhibition in cell viability across a dozen of various CRC cell lines, with IC50 values between 0.052 and 0.64 μM (Figure 1A and Table 1). Real-time cell analysis showed that the normalized cell index was inhibited by chaetocin in a dose-dependent manner in HCT116, LS174T and LIM1215 cells (Figure 1B), indicating that the cell proliferation of CRC cells was markedly restrained by chaetocin. Moreover, colony formation assay also revealed a decrease in both size and numbers in CRC colonies (Figure 1C). Collectively, our results demonstrated that chaetocin suppressed CRC cell growth in vitro. To determine whether chaetocin achieved this by inducing cell cycle arrest, cell cycle distribution analysis was carried out using flow cytometry. As shown in Figure 1D, after HCT116, LS174T and LIM1215 cells were treated with chaetocin for 12 h, the population of HCT116, LS174T and LIM1215 cells at the G₂/M phase increased substantially, indicating that chaetocin Inhibits CRC cell proliferation by prompting G2/M phase arrest.

Our previous studies have shown that chaetocin induces apoptosis in gastric cancer cell lines HGC-27 and AGS (Wen et al., 2019), whether the same would occur in CRC cells was investigated. After chaetocin treated, as shown in Figure 2A, flow cytometry exhibited that chaetocin increased the apoptotic cell populations in HCT116, LS174T and LIM1215 cells, and the apoptotic population increases as the duration of treatment lengthened. Because caspase pathway activation is crucial to apoptosis (Mandal et al., 2020), caspase family proteins were then assessed by western blot analysis. Chaetocin decreased the levels of the precursor forms of caspase-3, -8, and -9 but increased cleavage of those proteins in HCT116, LS174T and LIM1215 cells (Figure 2B). Consistently, chaetocin induced PARP cleavage which is the known downstream protein of caspase pathway (Figure 2B), illustrating that chaetocin induces CRC cell apoptosis which may depend on caspase activation. When apoptosis occurs, the integrity of the mitochondrial membrane will decrease, a phenomenon recognized as an indicator of apoptosis (Lindsay et al., 2011). As shown in Figure 2C, the integrity of the mitochondrial membrane was disrupted in HCT116, LS174T and LIM1215 cells after chaetocin treatment. To further confirm this, western blot was performed to evaluate the expression of several antiapoptotic proteins (BCL-2, BCL-XL, MCL-1 and XIAP) in CRC cells lines (Figure 2D). The decrease in antiapoptotic proteins expression further established that chaetocin induces apoptosis in CRC cells. Lastly, z-VAD-fmk, a pan-caspase inhibitor, was implemented as a pretreatment. After the caspase pathway was restrained by z-VAD-fmk (Supplementary Figure 2), the apoptotic cell population following treatment with chaetocin was significantly restored in CRC cells (Figure 2E), indicating that chaetocin-elicited apoptosis of CRC cells depends on caspase activation.

It was previously reported that chaetocin induces apoptosis through the accumulation of ROS in gastric cancer, glioma, and myeloma cells (Isham et al., 2007; Dixit et al., 2014; Wen et al., 2019). The fluorescent probe DCFH/DA was first used to validate that the ROS level rises after chaetocin treatment in HCT116 and LS174T cells (Figure 3A). Next, a ROS scavenger, NAC was used. Flow cytometry analysis showed that NAC almost abolished chaetocin-induced apoptosis in HCT116 and LS174T cells (Figure 3B). Moreover, PARP and caspase-3 cleavage induced by chaetocin were reversed when cells were cotreated with NAC (Figure 3C). These results suggest that Chaetocin-induced apoptosis in CRC cells relies on ROS accumulation.

It is well known that ROS induce cancer cell apoptosis by activating the JNK/c-Jun pathway (Aggarwal et al., 2019). To explore whether chaetocin-induced apoptosis take on the same ROS/JNK/c-Jun route, several experiments were carried out. First, western blot results showed a markedly increase in JNK and c-Jun phosphorylation after chaetocin treatment (Figure 4A), suggesting JNK/c-Jun pathway activation. SP600125, owing to its ability to suppress JNK/c-Jun pathway, was then used to prove that chaetocin-induced cell death depends on JNK/c-Jun activation. As shown by the results of flow cytometry and western blot, SP600125 partially restored chaetocin-induced apoptosis in HCT116 and LS174T cells (Figures 4B,C). We finally investigated if ROS accumulation induced by chaetocin relates to JNK/c-Jun activation. As shown in Figure 4D, NAC restored the abnormal increase in phosphorylation of JNK and c-Jun after chaetocin treatment. The above results suggest that Chaetocin induces cell death through the ROS/JNK/c-Jun axis.

We then explored whether chaetocin’s effect could stretch into 5-FU-resistant CRC cells, using cell lines that has already been established by our group (Wen et al., 2015; Wen et al., 2016). A similar and decent IC50 value of 0.054 μM was exhibited in the CCK8 cell viability assay (Figure 5A). Moreover, the same dose-dependent colony formation inhibition pattern could be observed in Figure 5B as well as the Increase in G2/M population as shown in Figure 5C. Flow cytometry and western blot results also certified that chaetocin induced apoptosis in HCT-15/5FU-R cells by caspase activation like that in normal CRC cells Figures 5D,E. These results illustrated that chaetocin could by pass 5-FU resistance in CRC cells.

To check that chaetocin induces cell death through the same pathway in 5-FU resistant strains of CRC, similar experiments were performed in HCT-15/5FU-R cells. Similar increase in ROS was detected in HCT-15/5FU-R cells (Figure 6A). NAC overturned the apoptotic effect induced by chaetocin (Figure 6B). The above results established that chaetocin induces accumulation of ROS and ultimately leads to apoptosis. Moreover, the increase in phosphorylation of JNK and c-Jun confirms that chaetocin leads to JNK/c-Jun pathway activation in HCT-15/5FU-R cells (Figure 6C). Also SP600125 was able to partially reverse the effect of chaetocin-induced apoptosis (Figures 6D,E). In addition, chaetocin was no longer able to induce JNK/c-Jun activation after the ROS in cells were scavenged by NAC (Figure 6F). The above results illustrated that chaetocin Induces apoptosis in HCT-15/5FU-R cells through the same ROS/JNK/c-Jun axis as in 5-FU-sensitive CRC cell lines.

In recent years, more and more studies have shown that CD47 could be one of the key immune checkpoint molecules in tumor therapy (Lian et al., 2019; Hayat et al., 2019). The present study showed that chaetocin not only reduced the expression of CD47 at mRNA levels (Figure 7A), but also inhibited the expression of total CD47 protein and membrane CD47 protein in CRC cells (Figures 7B,C). The highly expressed CD47 on the surface enables tumor cells to evade from immune surveillance of macrophages (Russ et al., 2018). Our results showed that macrophages had an enhanced phagocytosis of HCT116 cells which were pretreated with chaetocin (Figure 7D), indicating that chaetocin could reduce the expression of CD47 in CRC cells and enhance macrophages phagocytosis of CRC cells. As ROS/JNK/c-Jun involvement in chaetocin treatment was established in our study, whether the expression of CD47 was down-regulated through ROS/JNK/c-Jun signaling pathway was then investigated. When NAC was used to scavenge ROS in chaetocin-treated HCT116 cells, the inhibitory effects of CD47 at mRNA and protein levels were both partially reversed (Figures 7E,F), indicating that ROS accumulation down-regulated the expression of CD47 in CRC cells at transcriptional levels. Although SP600125 did not reverse the reduction of CD47 at mRNA level, it reversed the reduction of CD47 at protein level in HCT116 cells (Figures 7G,H), indicating a post-transcriptional regulation of JNK/c-Jun on CD47. All these results illustrated that chaetocin decreased the expression of CD47 at both mRNA and protein levels in CRC cells and enhanced macrophages phagocytosis.

The role of chaetocin in tumor growth was assessed in vivo using a nude mouse xenograft model by subcutaneously inoculating them with HCT116 cells. Mice were treated with vehicle and chaetocin (0.5 mg/kg) by intraperitoneal injection every day for 18 days. As shown in Figures 8A–C, the volume and weight of tumor in HCT116 xenografts were effectively suppressed in the chaetocin-treated group compared with control group, suggesting that chaetocin effects tumor proliferation both in vitro and in vivo. The protein levels of pro-caspase-3 were decreased in chaetocin-treated tumors, while p-JNK levels were increased (Figure 8D), indicating that chaetocin induces cell apoptosis like in vitro, undertaking the JNK/c-Jun pathway in CRC cell xenografts.