Sur-X, a novel peptide, kills colorectal cancer cells by targeting survivin-XIAP complex

The mRNA expression levels of IAPs in cancer tissues were compared with normal controls by Oncomine (https://​www.​oncomine.​org/​), an on-line cancer microarray database. The thresholds of p-value and fold change were set to 0.05 and 1.5, respectively.

GEPIA (Gene Expression Profiling Interactive Analysis) is an interactive web server for analyzing gene expression profiling of cancer and normal tissues from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) projects (http://​gepia.​cancer-pku.​cn/​). The expression profile of BIRC5/survivin in multiple cancers and all the IAPs expression in colorectal cancer and matched normal tissues were analyzed by GEPIA, parameters were set to default values.

The following primary antibodies were used for Western blot: antibodies against survivin (#2803), XIAP (#14334), Caspase 8 (#9746S), Caspase 9 (#9508S), Caspase 3 (#9662S), cleaved-Caspase 3 (#9661S), PARP (#9542 L), ubiquitin (#3933S), RIP1 (#3493S), phospho-RIP1 (Ser166) (#44590), RIP3 (#13526S), phospho-RIP3 (Ser227) (#93654), phospho-MLKL (Ser358) (#91689S), TAB1 (#3226), TAK1 (#5206), phospho-TAK1 (Thr184/187) (#4508) and GAPDH (#5174) were from Cell Signaling Technology; anti-Caspase 7 antibody (#sc-28,295) was purchased from Santa Cruz and anti-MLKL antibody (#184718) was purchased from and Abcam. Z-VAD-fmk (Sigma-Aldrich, V116) and Nec-1 s (Biovision, 2535–1) were used to reveal apoptotic and necroptotic cell death, respectively. Cycloheximide (CHX, C7698-5G) was purchased from Sigma-Aldrich. MG132 (S2619) was obtained from Selleckchem.

Human colorectal cancer cell lines HCT116, RKO, HCT15, and HT29 were obtained from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The human peritoneal mesothelial cell line HMrSV5 (SV5) was provided by Prof. Huimian Xu (Department of Surgical Oncology and General Surgery, The First Hospital of China Medical University). All cell lines were cultured with RPMI-1640 containing 10% fetal bovine serum (FBS, Gibco, Gaithersburg, MD, USA) and 1% penicillin/streptomycin. The short tandem repeats (STR) profile was used to authenticate all the cell lines. All cells were cultured for no longer than 2 months and negatively tested for mycoplasma contamination.

The peptide with the sequence derived from the XIAP-binding region (K15-M38) of survivin and the cell-penetrating sequence from HIV Tat protein added to its N-terminal, was named Sur-X. The sequence of Sur-X is YGRKKRRQRRRKDHRISTFKNWPFLEGCACTPERM-COOH. A negative control peptide is YGRKKRRQRRRKDDGNYKTRAEVKFEGDTLVNRIE-COOH, and named Con. Peptides Sur-X and Con, as well as FITC labeled peptides FITC-Sur-X and FITC-Con were all synthesized by Synpeptide Co. Ltd. (Nanjing, China). DMSO was used to dissolve these peptides immediately before use. Final concentrations of DMSO in experiments were < 0.4%.

Cells were seeded into 2 cm glass-bottom cell culture dish (NEST, 801001) and treated by 10 μM FITC-Sur-X or FITC-Con after attachment. Hoechst 33342 (BD Biosciences, 561,908) was used to stain nuclei. After 30 min, the culture medium was discarded and cells were washed twice using ice-cold PBS and inspected under an Olympus FluoView FV1000 laser scanning confocal microscope.

When the confluency reached 90% in a 10-cm dish, cells were lysed in 1% Triton lysis buffer (1% Triton X-100, 50 mM Tris-Cl pH 7.4, 150 mM NaCl, 10 mM EDTA, 100 mM NaF, 1 mM Na₃VO₄, 1% protease inhibitor cocktail) on ice for 5 min and collected in a 1.5 ml EP tube. Retaining 10% of cell lysate as input, the remaining was subjected to co-immunoprecipitation assay and Western blot analysis, which were performed as described in our previous studies [24, 25]. Image J (NIH, USA) was used to perform densitometric analysis.

3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay was used to detect cell viability as previously described [24].

Cells were seeded into 12-well plates and treated by 10 μM Sur-X or Con. Ten days later, cells were stained by Wright-Giemsa after 75% ethanol-fixation. The colonies in each well were counted in five fields under an optical microscope.

Trizol-chloroform was used to extract total RNA and qRT-PCR was performed as previously described [26]. Primers were listed in Table S1.

HCT116 and RKO cells were seeded at the density of 4000 cells/well in 100 μl culture medium into Nunclon 96-Well flat white plate (Thermo Fisher Scientific). After cells were attached, 2× Sur-X of finally detected concentrations and 2× detection reagents in RealTime-Glo™ Annexin V Apoptosis and Necrosis Assay (Promega, JA1011) were mixed in 100 μl pre-warming culture medium and the mixture was added to each well. Relative luminescence units (RLU) and relative fluorescence units (RFU) were recorded overtime up to 6 h by the Spark multimode microplate reader (Tecan Trading AG, Switzerland).

The activity of Caspase 3 was measured according to the instruction of Caspase 3 Activity Assay Kit (Cell Signaling Technology, #5723). RFU with excitation at 380 nm and emission at 440 nm was detected by Thermo Scientific Varioskan Flash.

HCT116 cells were incubated with 10 μM MG132 for 8 h and then treated by 10 μM Sur-X for 1 h. Subsequently, immunoprecipitation was implemented as previously described. After SDS-PAGE, the immune-precipitates were probed with indicated antibodies including anti-ubiquitin antibody. The experiment was performed for three times independently.

After indicated treatment, HCT116 and RKO cells were collected and prepared for flow cytometry analysis by using FITC Annexin V Apoptosis Detection Kit with 7-AAD (Biolegend, 640,922), according to the staining procedure recommended by the manufacture. BD Accuri C6 Flow Cytometer was used for the detection and analysis of samples.

To predict the binding mode of XIAP and Sur-X, the structure of XIAP was constructed based on the crystal structures of BIR1, BIR2, BIR3 domains (PDB codes: 4OXC, 4J3Y, and 1G73) and the missing residues were filled using YASARA. The assembled XIAP structure was then subject to three step-wise rounds of energy minimization to relax the loops that connect different domains. For Sur-X, we extracted corresponding atomic coordinates (residues number 15 to 38) from the crystal structure of survivin (PDB code: 1F3H). Starting from this conformer, a total of 400 ns accelerated molecular dynamics (aMD) simulations were carried out and the center of the highest-populated cluster obtained from aMD trajectories was selected as the representative low-energy conformational state of the peptide [27, 28]. Afterwards, the initial XIAP-peptide complexes were created by using ZDOCK [29]. Finally, the binding poses were clustered into 27 complexes and selected for further molecular dynamics (MD) simulations.

All MD simulations were performed by using Amber 16 program (University of California, San Francisco) on GPUs. The Amber99sb force field was used for all systems [30]. For each binding pose, the complex structure was solvated in a cubic periodic box of TIP3P water with at least 10 Å distance from the complex surface, followed by the addition of 0.15 M NaCl ions to neutralize the system. To relax the whole system, the starting structure was then subject to energy minimization and two rounds of 500-ps constant volume (NVT) and 500-ps constant pressure (NPT) equilibration with position restraints on the solute heavy atoms. Subsequently, the production NPT simulations were conducted for up to 200 ns for each complex system. During the production runs, all bonds involving hydrogen were constrained using the SHAKE algorithm to enable an integration time-step of 2 fs. The non-bonded interaction was cut off at 1.0 nm, whereas the electrostatic interaction beyond that was treated using the particle mesh Ewald (PME) method. The Langevin Thermostat was applied to maintain the temperature at 300 K, and the Berendsen Barostat was used to maintain the pressure at 1.0 bar. The trajectories of the production runs were saved at 8 ps intervals.

We collected 675,000 coordinates from all of the 5.4 μs production runs and performed our analysis on the basis of these snapshots. The interaction residues between XIAP and Sur-X were determined by the following criterion: if the shortest distance between any heavy atoms from two residues belonging to two proteins was less than 4.5 Å, then these two residues from two proteins were considered to form an interaction residue pair. By counting the highest-frequent pairs appeared during the 60–100, 100–160, and 160–200 ns trajectories of each simulation, we determined the interaction pairs shared between the three time periods as the interface residues between XIAP and Sur-X.

According to the manufacturer’s instruction, Human TNF-α ELISA kit (R & D Systems, DTA00D) was used to measure the TNF-α in culture medium secreted by colorectal cancer cells.

TAB1 plasmid and the vector pcDNA3.1(+) were purchased from Obio Technology Corp., Ltd. (Shanghai, China). SiRNAs of TAK1 and negative control (NC) were obtained from Beijing ViewSolid Biotech, the sequences are as follows: 5′-GGUGCUGAACCAUUGCCAUTT-3′ for si#1 of TAK1; 5′-GCAACCCAAAGCGCUAAUUTT-3′ for si#2 of TAK1; 5′-UUCUCCGAACGUGUCACGUTT-3′ for NC. Cells were transfected as previously described [25].

A total of 10 four-week old female BALB/c nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) and fed at a specific pathogen-free environment in Animal Laboratory Unit of China Medical University, with the approval (2019088) of Institutional Review Board of China Medical University for all in vivo experiments. HCT116 cells (5 × 10⁶) were injected subcutaneously near the right scapula of mice. When the average tumor volume reached 50–75 mm³, mice were randomly divided into two groups (n = 5) and received intravenous injection of 50 mg/kg Sur-X or Con daily for a total of 14 times since then. Tumor size and body weight were measured every other day. The volume of each tumor V was calculated using V = 1/2 (length × width²). Mice were killed by cervical dislocation according to the protocol filed with the Guidance of Institutional Animal Care and Use Committee of China Medical University.

Tumor tissues were fixed by formalin, embedded by paraffin and prepared for staining with haematoxylin and eosin (HE) and antibodies of survivin, XIAP, and MLKL as described in our previous studies [31]. The staining was evaluated by scanning the entire tissue specimen under low magnification (× 10) and confirmed under high magnification (× 20 and × 40). Both staining intensity and staining area were used to classify the expression of proteins, with staining intensity scored as 0 (no), 1 (low), 2 (intermediate), and 3 (high) points and staining area scored as 0 (≤5%), 1 (5–25%), 2 (25–50%), 3 (50–75%) and 4 (> 75%) points, respectively. Histoscore was calculated as histoscore = staining intensity × staining area. Two pathologists were responsible for determining the final histoscore independently. The expression of protein with a histoscore of 0, 1–4 points and 6–12 points were defined as negative (−), weak positive (+) and strong positive (++), respectively and evaluated under 5 randomly selected, non-overlapping fields from the stained sections.

One Step TUNEL Apoptosis Assay Kit (Beyotime Biotechnology, China, C1088) was used to detect the apoptosis in xenograft tumors according to the manufacture’s instruction. Fluorescence microscope BX53 (Olympus, Japan) was used to visualize the stained tissue sections.

SPSS Version 16.0 (SPSS Inc., Chicago, IL) was used to analyze the results of experiments which were conducted in triplicate and presented as mean ± standard deviation (SD). The graphics were generated using GraphPad 6.0 (GraphPad Software, USA). Student’s t-test and one-way ANOVA were used to analyze the differences between two independent groups and multiple groups, respectively; Chi-square test was used to assess the differences of categorical data; p < 0.05 was considered statistically significant.

Online databases Oncomine and GEPIA were used to compare the expression of IAPs (NAIP, BIRC2, BIRC3, XIAP, BIRC5/survivin, BIRC6 and BIRC7, no data on BIRC8 was available) between colorectal cancer and normal tissues. As the most widely studied member of IAPs, survivin was significantly overexpressed in most cancers available in Oncomine including colorectal cancer, which was validated by the analysis of survivin expression profile in multiple cancers by GEPIA (Fig. 1a and Figure S1A). Furthermore, the analysis by GEPIA suggested that among other IAPs apart from survivin, the expression of XIAP was also higher in colorectal cancer tissues, although with no statistical significance (Fig. 1b and Figure S1B). These data indicated that survivin and XIAP are overexpressed in colorectal cancer and are potential therapeutic targets for colorectal cancer treatment.

It has been reported that survivin and XIAP are able to form an IAP-IAP complex to stabilize both of them and synergistically antagonize the activity of Caspase 9 in cancer cells [23]. As shown in Fig. 1c, we performed co-immunoprecipitation of survivin and XIAP in colorectal cancer cells HCT116 and RKO, and found that survivin-immunocomplexes contained XIAP, and XIAP-immunocomplexes also contained survivin, demonstrating that these two proteins can form survivin-XIAP complex in colorectal cancer cells. Therefore, we inferred that blocking the interactions between survivin and XIAP might be a promising therapeutic strategy in colorectal cancer.

According to the finding of a previous study that the minimal XIAP-interacting region of survivin comprised residues K15-M38 [32], we designed and synthesized a peptide that targets the survivin-XIAP complex, Sur-X, which overlapped with residues of survivin from K15 to M38, and the TAT sequence was added to the N-terminal of the peptide to enhance its cell membrane permeability (Fig. 1d). Peptide Con consisted of the N-terminal TAT sequence and a random 24-amino acid sequence, used as a negative control.

To investigate the cell membrane permeability of the peptides, Sur-X and Con were labeled with FITC and the internalization of FITC-Sur-X and FITC-Con was observed within 30 min (Fig. 1e). Moreover, it was confirmed that Sur-X indeed disrupted the formation of survivin-XIAP complex in colorectal cancer cells. As shown in the left panel of Fig. 1f, less survivin was co-immunoprecipitated with XIAP in colorectal cancer cells treated by Sur-X for 1 h; similarly, the right panel indicated a significant reduction of XIAP in survivin-immunocomplexes in Sur-X-treated cells. These results suggested that Sur-X can interfere with the formation of survivin-XIAP complex in colorectal cancer cells and may have anticancer effects.

To evaluate the anticancer activity of Sur-X in colorectal cancer cells, four human colorectal cancer cell lines (HCT116, RKO, HCT15, and HT29) were exposed to a series of concentrations of Sur-X for 1, 3, 6, and 24 h, and their viabilities were assessed by MTT assay. Of note, viabilities of these four cell lines were significantly decreased in a dose-dependent manner with the treatment of Sur-X (Fig. 2a-b and Figure S2A-B). However, no inhibitory effect of Sur-X was observed in human peritoneal mesothelial cell line SV5 (Fig. 2c and Figure S2C). As the negative control, Con was ineffective in both colorectal cancer cell lines and normal cell line SV5 (Fig. 2a-c and Figure S2A-C). Similarly, the colony formation assay also confirmed the antitumor activity of Sur-X in colorectal cancer cells (Fig. 2d). Then, to explore whether the specific anticancer activity of Sur-X was dependent on survivin and XIAP, their expression levels in all the five cell lines were detected by Western blot analysis and it was found that both survivin and XIAP were highly expressed in colorectal cancer cells, but not detected in SV5 (Fig. 2e).

Since the interactions between survivin and XIAP were decreased significantly upon the treatment of Sur-X for only 1 h (Fig. 1f), we speculated that Sur-X might induce apoptosis in colorectal cancer cells by destabilizing survivin and XIAP. Then, real-time apoptosis detection was performed to evaluate the apoptosis-inducing ability of Sur-X in HCT116 cells. As shown in Fig. 3a, a time-and dose-dependent significant increase in apoptotic signals (RLU) was observed in cells treated by Sur-X at indicated concentrations, except for 2.5 μM. Of note, Sur-X-induced apoptosis reached a plateau around 3 h at most concentrations (Fig. 3a), which was consistent with the finding that Sur-X rapidly inhibited the viability of colorectal cancer cells (Fig. 2a-b and Figure S2A-B). In addition, compared with untreated cells, significant increase in Caspase 3 activity was observed in Sur-X-treated cells (Fig. 3b). Further, the effects of Sur-X on the expression levels of apoptosis-related proteins were evaluated by Western blot. As shown in Fig. 3c, the cleavages of Caspase 9, 3, 7 and PARP were increased in both HCT116 and RKO cells treated with Sur-X, however, no activation of Caspase 8 was observed. Additionally, as expected, Sur-X decreased the expression of survivin and XIAP (Fig. 3c). Next, it was proved that Sur-X had no significant inhibitory effect on the mRNA expression of survivin and XIAP (Fig. 3d). However, when protein synthesis in colorectal cancer cells was blocked by CHX, Sur-X was found to promote the degradation of survivin and XIAP (Fig. 3e). Moreover, the ubiquitination levels of both survivin and XIAP were enhanced in colorectal cancer cells with the treatment of Sur-X (Fig. 3f). Additionally, it was noteworthy that Sur-X did not interfere with XIAP-Caspase 9 interactions (Figure S3). Together, these results indicated that the disruption of survivin-XIAP complex by Sur-X caused ubiquitination-mediated degradation of survivin and XIAP, and subsequently promoted Caspase 9-dependent intrinsic apoptosis in colorectal cancer cells.

Moreover, what is worth noting is that as shown in Fig. 3g and h, pre-treatment with z-VAD, a pan-caspase inhibitor, only partially rescued Sur-X-induced cell death in HCT116 and RKO cells, which indicated that apart from apoptosis, Sur-X might also induce other forms of programmed cell death.

To study whether Sur-X induced necroptosis in colorectal cancer cells, we first evaluated the effects of Sur-X on cell morphology by Giemsa staining. Of note, different from the classic apoptotic morphology (condensed chromatin and intact cell membrane) induced by adriamycin, necrotic morphology was observed in Sur-X-treated cells (Fig. 4a). In addition, real-time necrosis detection showed significant increase in RFU in cells treated by Sur-X in a dose- and time-dependent manner, and flow cytometry analysis presented a rapid increase of 7-AAD-positive cells in Sur-X-treated cells, which also suggested that Sur-X might induce necroptosis (Fig. 4b and c). Furthermore, expression levels of necroptosis-related proteins including Receptor-Interacting Protein 1 (RIP1), RIP3, Mixed Lineage Kinase Domain-Like (MLKL) and their phosphorylation in colorectal cancer cells following treatment with Sur-X were examined by Western blot analysis. As shown in Fig. 4d, after treatment with Sur-X for 30 min, p-RIP1, p-RIP3, and p-MLKL were all significantly increased. More importantly, Nec-1 s, a necroptosis inhibitor, partially attenuated Sur-X-induced cell death mainly by significantly reducing 7-AAD-positive cells, which also confirmed the contribution of necroptosis to the anticancer activity of Sur-X (Fig. 4e and f). In addition, rapid increases in necrotic signal (RFU) and apoptotic signal (RLU) were detected almost simultaneously in Sur-X-treated cells by real-time apoptosis and necrosis analysis (Fig. 4g and Figure S4A), which further distinguished necroptosis from secondary necrosis, the latter is manifested by a significant increase in RFU after the RLU has reached a plateau. Taken together, these data suggested that necroptosis also contributed to Sur-X-induced cell death.

Since survivin could bind to the BIR1 and BIR3 domains of XIAP [23], we hypothesized that Sur-X might promote necroptosis by binding to XIAP to block the interactions between XIAP and other necroptosis-related essential proteins. Therefore, MD simulations were used to characterize the binding mode of Sur-X with XIAP. As shown in Table 1, 11 residues from Sur-X and 10 residues from XIAP formed 17 interaction pairs (the interactions between polar amino acids are mainly hydrogen bonds and those between non-polar amino acids are mainly hydrophobic interactions). Among the 10 residues from XIAP contributing to its binding with Sur-X, eight (L30, A34, S38, S45, D71, K85, N89, and F92) were located in the BIR1 domain (E26-I93). Thus, we speculated that Sur-X might interfere with the binding of other proteins to the BIR1 domain of XIAP.

Interestingly, by searching for “XIAP BIR1” in the PDB database, it was found that the BIR1 of XIAP could interact with TAB1 (Fig. 5a). Besides, a previous study reported that XIAP-TAB1 interactions were crucial for the activation of TAK1, an endogenous inhibitor of necroptosis [33, 34]. Thus, it was plausible that Sur-X might disrupt the XIAP-TAB1 interactions to relieve the inhibition of TAK1 on necroptosis. To this end, we first confirmed the formation of the XIAP-TAB1-TAK1 complex in colorectal cancer cells by the detection of XIAP and TAK1 in co-immunoprecipitation with TAB1 (Fig. 5b). Then, it was demonstrated that the interactions between XIAP and TAB1 were significantly decreased in cells with the treatment of Sur-X for 1 h (Fig. 5c). Taken together, this set of experiments suggested that Sur-X can block the interactions between XIAP and TAB1. Additionally, since it was found that the protein levels of both TAK1 and p-TAK1 were significantly decreased in Sur-X treated cells, whereas no change in TAB1 was observed (Fig. 5d), we speculated that disruption of XIAP-TAB1 interaction by Sur-X might promote the degradation of TAK1. As expected, compared with Con, Sur-X significantly accelerated the degradation of TAK1 (Fig. 5e).

To further identify the role of TAB1 and TAK1 in Sur-X-induced necroptosis, expressions of TAK1 and TAB1 in colorectal cancer cells were altered alone or in combination. Of note, knockdown of TAK1 increased the levels of p-RIP1 and p-MLKL and enhanced the anticancer activity of Sur-X in colorectal cancer cells (Fig. 5f-g and Figure S4B-C). However, TAB1 overexpression (TAB1-OE) attenuated the inhibition effects of Sur-X on colorectal cancer cells and decreased the expression of p-MLKL, which further confirmed the disruption of XIAP-TAB1 interactions by Sur-X (Fig. 5h-i and Figure S4D-E). Moreover, knockdown of TAK1 restored the anticancer effect of Sur-X and p-MLKL expression in TAB1-OE cells (Fig. 5h-i and Figure S4D-E). All these results indicated that Sur-X promoted necroptosis by interfering with the interactions between XIAP and TAB1 and subsequently destabilizing TAK1 to alleviate the inhibition of TAK1 on necroptosis in colorectal cancer cells.

In addition, considering the contribution of XIAP-TAB1-TAK1 complex to NF-κB activation [33], the effects of Sur-X on NF-κB signaling pathway were detected. As shown in Figure S5A, after treated by Sur-X for 30 min or1 h, the expression levels of p-IκB and p-P65 in colorectal cancer cells showed a transient increase. Interestingly, the most significant increase of TNF-α secreted by colorectal cancer cells was observed after the treatment of Sur-X for 30 min (Figure S5B). Thus, it was plausible that Sur-X induced a rapid and transient secretion of TNF-α by colorectal cancer cells and subsequently activated NF-κB signaling pathway for a short time.

To validate the anticancer efficacy and safety of Sur-X in vivo, subcutaneous xenograft mouse model was established and received intravenous injection of Sur-X and Con (Fig. 6a). Of note, significant inhibition on the growth of tumors was observed in mice with the treatment of Sur-X (Fig. 6b). Importantly, no significant difference in body weight was found between Sur-X-treated mice and Con group, and no toxicity of Sur-X to vital organs such as heart, liver and kidneys was observed (Fig. 6c and d). Furthermore, the decreased expression of survivin and XIAP and enhanced staining of TUNEL were observed in tumor tissues from mice treated by Sur-X, which indicated the induction of apoptosis by Sur-X in vivo (Fig. 6e, f and Figure S6A). In addition, the in vivo pro-necroptosis effect of Sur-X was confirmed by the extensive necrosis and the reduced MLKL expression in tumor tissues from Sur-X-treated mice (Figure S6B, Fig. 6g and h).

Altogether, in vitro and in vivo data indicated that Sur-X exerted specific and potent anticancer effects by promoting apoptosis and necroptosis in colorectal cancer cells.