Introduction

Cancer is a widespread disease and considered as one of the leading causes of death worldwide. Chemotherapy and radiotherapy are the main approaches for cancer therapy in addition to surgery. However, due to the lack of tumor specificity, they kill both cancer cells and normal cells, eliciting severely toxic side effects1. Hence, ideal agents for cancer treatment are those can selectively kill cancer cells without harming normal cells. So TRAIL is a promising candidate for cancer therapy, as it can selectively kill tumor cells but with no or little harm toward normal cells. And clinical trials involving recombinant rhTRAIL also showed good outcomes without obvious side effects.

TRAIL or Apo2L, a type II transmembrane protein, with an extracellular region forming a soluble molecular when cleaved by protease, is a member of the tumor necrosis factor (TNF) family. Compared with other members of this family, TRAIL can induce apoptosis of various transformed cells and xenografts, leaving normal cells unharmed in vitro2,3,4. In vivo, TRAIL differs from TNF-α and FasL. Besides inducing apoptosis of cancer cells, TNF causes severe inflammatory responses and FasL causes liver damage, when administered systemically5. However, TRAIL activates nuclear factor κB (NF-κB) very weakly and it is unlikely to initiate inflammatory cascades when administered systemically. P53, which is considered as a critical factor of cancer therapies involved in chemotherapeutic agents or radiation6, is mutated in at least 50% of all human cancers7. However, TRAIL-induced apoptosis does not depend on p53 status, for this reason, TRAIL can be used to overcome chemotherapeautic or radiotherapeautic resistance caused by p53 dysfunction8. Therefore, TRAIL is a promising candidate of cancer therapy. Since its discovery in 1995, TRAIL has been intensely investigated. However, the applications of TRAIL were impeded because certain TRAIL preparations were toxic against primary human hepatocytes via a caspase-dependent mechanism that involved the activation of the extrinsic death pathway7,9. Further studies have suggested that the use of native TRAIL, not tagged TRAIL, can prevent such toxicities10.

Two membrane-associated death receptors, DR4 or/and DR5, are required in TRAIL induced-apoptosis. The interaction of TRAIL with DR4 or DR5 is an initial step of TRAIL-induced apoptosis11,12. The binding of trimeric TRAIL to death receptors results in receptors homotrimerization13,14. Activated death receptors recruit Fas-associated DD (FADD), which is an adaptor protein that can directly bind to the intracellular death domain of death receptors. Through the interactions of DED of FADD and pro-caspase-8/10, the pro-caspase-8/10 is recruited, thereby forming the death-inducing signaling complex (DISC)15,16,17. In DISC, pro-caspase-8/10 is auto-cleaved to generate an active form. According to the need of the mitochondrial pathway to trigger TRAIL-induced apoptosis completely, cells are divided in two groups, namely, type I cells and type II cells15. In type I cells, activated caspase-8 is sufficient to activate downstream effector caspases such as caspase-3 and caspase-7 to execute apoptosis18,19,20. In type II cells, activated caspase-8 is not sufficient to trigger apoptosis and mitochondria pathway is needed to amplify apoptotic signal20,21,22.

Sensitivity to TRAIL-induced apoptosis is a key factor influencing efficacy of cancer treatment. However, it has been reported that not all the cancer cells are sensitive to TRAIL-induced apoptosis due to the acquired or inherent resistance23. The resistance of cancer cells has been attributed to dysfunction of different steps in TRAIL-induced apoptosis pathway or elevation of survival signals24. Studies have indicated that the resistance appeared to be mediated by the following factors: mutation, modification and location of death receptors; increased expression of caspase inhibitors such as c-FLIP, X-linked inhibitor of apoptosis protein (XIAP), cellular inhibitor of apoptosis protein (cIAP) and surviving; or increased expression of Bcl-2 protein family, including Bcl-2 and Bcl-XL25,26,27. The activation of other survival pathways, including AKT and NF-κB pathways, also leads to cell survival and chemoresistance28. However, the resistant mechanisms are not fully understood now. Studies on intracellular mechanisms that mediate the resistance of TRAIL-induced apoptosis may help develop TRAIL-based approaches for cancer therapy. The combinations of TRAIL with chemotherapy or radiotherapy are used to overcome resistance of some cancers, in which monotherapy with TRAIL, chemical agents or radiation is ineffective.

Annexin V is a calcium phospholipid-binding protein that belongs to the Annexin superfamily. It is widely expressed in skeletal, cardiac, smooth musle, Leyding cells, endothelia, chondrocytes and some neurons29,30,31,32,33. Previous studies have revealed that Annexin V is involved in various extracellular and intracellular processes, including blood coagulation, signal transduction, anti-inflammatory processes, membrane trafficking and ion channel activity34,35,36. Besides, a recent study has revealed that Annexin V competes with tumor-derived microvesicles containing activated EGFR to bind to PS of tumor-derived vascular endothelial cells and inhibits tumor angiogenesis37.

In this study, a chimeric protein Annexin V-TRAIL (designated as TP8) with higher efficacy than TRAIL was developed, which selectively induced various kinds of tumor cell apoptosis in vitro and in vivo. In vitro, the EC50 of TP8 was lower. TP8 exhibited long-term effects in addition to rapid induction of apoptosis because colony growth was suppressed by TP8. In vivo, TP8 effectively suppressed tumor growth in mouse xenografts involving multiple cancer cell types including A549, Colo205 and Bel7402. Besides, tumor doubling time and tumor growth delay time were also evidently prolonged by TP8's treatment, but not by PBS, Annexin V, TRAIL or Annexin V + TRAIL. All the results suggested TP8 might be a potential candidate to treat TRAIL-resistant cancer. This study provides a new rational strategy to treat cancer.

Results

Annexin V endowed the TRAIL fused-protein TP8 with higher polymerization ability

After purification, the purity of TP8 final product was over 98% (Figure 1A). The purified recombinant protein TP8 was comprised of 31.69% monomer, 38.70% dimer and 29.61% trimer. However, 70.14% monomer, 27.25% dimer and 2.62% trimer were present in TRAIL alone or TRAIL + Annexin V (Figure 1B and 1C). The different trimer and dimer proportions between TRAIL and TP8 indicated that these molecules might exhibit different bioactivities, as one previous study has reported that the apoptosis-inducing activity of TRAIL is related to its ability to form homotrimers38. The presence of Annexin V endowed the fused TRAIL protein with higher trimer and dimer formation ability, indicating that TP8 might have enhanced apoptosis-inducing ability.

Figure 1
figure 1

SDS-PAGE analysis for purified TP8, TRAIL or Annexin V + TRAIL.

(A) TP8 analyzed under reducing condition. Lane 1: molecular weight markers; Lane 2: purified TP8. Full-length gel is presented in Supplementary Figure 1. (B) TRAIL or Annexin V + TRAIL analyzed under non-reducing condition. Lane 1: molecular weight markers; Lane 2: Annexin V; Lane 3: TRAIL; Lane 4: Annexin V + TRAIL. Full-length gel is presented in Supplementary Figure 2. (C) TP8 analyzed under non-reducing condition. Full-length gel is presented in Supplementary Figure 3.

TP8 enhanced the antitumor activity of TRAIL but maintained tumor specificity of TRAIL in vitro

Human MDA-MB-231, Colo-205, A549, Bel7402, MCF-7, HeLa, HepG2, U937, H446, H1229 and PLC cells, treated with varying concentrations of TRAIL or TP8 for the indicated time, were assessed by MTT assay to determine the cytotoxicity of TRAIL and TP8. For TRAIL-sensitive MDA-MB-231 and Colo-205 cells, the EC50 of TRAIL was 5.19 and 2.74-fold that of TP8, respectively. In TRAIL moderate-sensitive or TRAIL-resistant cancer cells, TP8 also exhibited severe cytotoxicity even at low concentrations (Table 1). In contrast to wild-type TRAIL, TP8 induced significant and extensive apoptosis (>50%) rapidly, when they were applied at the same concentration. For example, the robust apoptosis of Colo-205 and MDA-MB-231 cells was induced at 2 hours, whereas the apoptosis of A549 cells was induced at 6 hours (Figure 2D, 2E and 2F). Moreover, TP8 was more effective than RGD-TRAIL, a TRAIL derivative with the ability to target TRAIL to tumor vasculature, which was reported to exhibit 3–5 fold higher apoptosis-inducting activity than TRAIL in vivo39 (Figure 2C). Although TP8 greatly enhanced the apoptosis-inducing activity of TRAIL, tumor specificity remained unchanged. 293 T cells were used to evaluate the cytotoxicity of TP8 toward normal cells. The results of Fig. 3D showed that TP8 had no effects on 293 T cells, as the apoptosis did not increase with the increase of TP8 concentration, even when the cells were treated with 800 ng/ml TP8 for 72 hours, no significant apoptosis occurred. That's to say, TP8 exerted no cytotoxicity on normal 293 T cells (Figure 3D).

Table 1 The EC50 (nM) of TRAIL and TP8 on different tumor cell lines
Figure 2
figure 2

TP8 induced apoptosis more rapidly than TRAIL and with better effects than TRAIL and RGD-TRAIL.

(A) Kinetics of TP8-induced apoptosis in A549 cells. Cells were incubated with TP8 (600 ng/ml) for indicated time. (B) Dose dependency of TP8 in A549 cells. Cells were incubated with TP8 for 6 hours at the concentrations indicated. (C) Antitumor effect comparisons of TRAIL, RGD-TRAIL and TP8 toward A549 cells. TP8 induced intensive apoptosis of (D) Colo-205, (E) MDA-MB-231 and (F) A549 cells more rapidly than TRAIL. Columns, means of three independent experiments; bars ± SD.

Figure 3
figure 3

The in vitro cytotoxicity of TP8 towards A549 cells and 293 T cells.

(A) A549 cells were treated with 10.74 nM Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours and cell death was analyzed using flow cytometer. (B) TEM images for A549 cells morphology after treatment with 10.74 nM Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours. (C) Colony-forming assay for A549 cells after treatment with 1.79 nM Annexin V, TRAIL, Annexin V + TRAIL or TP8 for indicated time. (D) 293 T cells were treated with 10, 50, 100, 200, 400, 600 and 800 ng/ml TP8 TP8 for 72 hours. No significant apoptosis was observed. Columns, means of three independent experiments; bars ± SD.

TP8 induced apoptosis and reduced colony growth of A549 cells

To investigate the apoptosis-inducing activities of Anneixn V, TRAIL, Annexin V + TRAIL and TP8, A549 cells were incubated with PBS or the aforementioned proteins for 6 hours. Consistent with cell viability results, only TP8 induced robust apoptosis of A549 cells, whereas Annexin V, TRAIL or Annexin V + TRAIL were ineffective (Figure 3A). To confirm these results, we investigated the changes of cellular morphology. A549 cells, treated with PBS, Annexin V, TRAIL, Annexin V + TRAIL or TP8, were examined by TEM. Only the cells treated with TP8 exhibited apoptotic morphology (Figure 3B).

To assess whether or not TP8 affects the long-term growth of A549 cells, we examined the colony formation of cells treated with PBS, Annexin V, TRAIL, Annexin V + TRAIL or TP8. The results showed that the colony growth of A549 cells was obviously reduced only when TP8 was applied (Figure 3C). All the results above suggested that TP8 exhibited both short and long-term effects.

Time- and dose-dependence of TP8-induced A549 cell apoptosis

To investigate the effects of treatment time and TP8 dose on the apoptosis of A549 cells, we treated the cells with various concentrations of TP8 or for various treatment periods. Apoptosis increased as the treatment period was extended. The apoptotic rate reached the maximum when the cells were treated for 6 hours (Figure 2A). After 6 hours, the apoptotic rate increased further as treatment period was prolonged, but this increase was not significant. During treatment with different concentrations of TP8 for the same duration, apoptosis rose with the increase of TP8 dosage (Figure 2B).

TP8 induced apoptosis in a caspase-dependent manner

The binding ability of TP8 to death receptors was not enhanced, compared with that of TRAIL or Annexin V + TRAIL (Figure 4A). The expression levels of DR4 and DR5 were not altered by TP8, compared with PBS, Annexin V, TRAIL or Annexin V + TRAIL (Figure 4B, 4C and 4D). However, western blot results showed that PARP, the substrate of caspase-3, was significantly activated. Besides, both the results of western blot and enzymatic activity assay suggested that caspase-8 and caspase-3 were evidently activated only after the A549 cells were treated with TP8 (Figure 4B, 4E and 4F). The pretreatment with Z-IETD-FMK, a caspase-8 inhibitor, inhibited TP8-induced apoptosis, but the pretreatment with caspase-9 inhibitor Z-LEHK-FMK did not reduce apoptosis induced by TP8 significantly (Figure 4G). When Cys230 of TRAIL domain of TP8 was mutated to Gly, the protein TP8-V could not induce apoptosis anymore, indicating that the activity of TP8 was dependent on the TRAIL domain (Figure 4H). Hence, all the results above demonstrated that TP8 induced apoptosis in the same way as TRAIL, indicating that TP8 would not cause additional cytotoxicity than TRAIL.

Figure 4
figure 4

TP8 induced cancer cells death in the same manner as TRAIL.

(A) Flow cytometer assay for the binding abilities of Annexin V, TRAIL, Annexin V + TRAIL and TP8 with A549 cells. (B) Western blotting detection for DR4, DR5, pro-caspase-8, pro-caspase-9, pro-caspase-3 and PARP. Full-length blots are presented in Supplementary Figure 4. (C) DR4 and (D) DR5 expression levels were analyzed using flow cytometer. (E) Caspase-8 activation determined with caspase-8 activity assay kits and (F) caspase-9 activation determined with caspse-9 activity assay kits. (Columns, means of three independent experiments; bars ± SD, **p < 0.01 compared with each corresponding treatment). (G) Caspase inhibitors study for the roles of caspases in TP8 induced apoptosis (Columns, means of three independent experiments; bars ± SD, **p < 0.01 compared with TP8 treatment). (H) Mutation of Cys230 of TRAIL in TP8 attenuated TP8 induced apoptosis. (Columns, means of three independent experiments; bars ± SD).

Administration of TP8 significantly suppressed tumor growth and prolonged tumor doubling time and tumor growth delay time

We confirmed the aforementioned results in vivo. A549, Colo205 and Bel7402 xenografts, which represent non-small cell lung cancer, colon cancer and liver cancer, respectively, were used to construct tumor models in athymic nude mice. In these three tumor models, tumor growth was effectively suppressed by TP8, but not by equal dosage of Annexin V, TRAIL or Annexin V + TRAIL. TP8 also retarded tumor growth in a dose-dependent manner, because the inhibitory efficacy was enhanced as dosage increased (Figure 5A, Figure 6A and 6C). Additionally, the tumor volume in mice treated with TP8 was significantly smaller than that of the controls (Figure 5A, Figure 6A and 6C). Furthermore, tumor growth of the TP8 group was slower than that of the controls. In mice bearing A549 xenograft, tumor doubling time was prolonged from 5.3 days (CI, 4.9–5.9 days) in mice receiving PBS, 6.1 day (CI, 5.6–6.8 days) in mice receiving Annexin V, 6.1 day (CI, 5.6–6.7 days) in mice receiving TRAIL, or 6.0 days (CI, 5.5–6.6 days) in mice receiving Annexin V + TRAIL, to 7.4 days (CI, 6.6–8.4 days) in mice receiving TP8 (12.5 mg/kg) (Figure 5B and Table 2). In mice bearing Bel7402 xenograft, tumor doubling time was prolonged from 7.3 days (CI, 7.0–7.7 days) in mice receiving PBS, 7.8 day (CI, 7.4–8.2 days) in mice receiving Annexin V, 8.0 days (CI, 7.7–8.4 days) in mice receiving TRAIL, or 7.6 days (CI, 7.3–8.0 days) in mice receiving Annexin V + TRAIL, to 9.3 days (CI, 8.9–9.6 days) in mice receiving TP8 (12.5 mg/kg) (Figure 6B and Table 2). In mice bearing Colo-205 xenograft, tumor doubling time was prolonged from 10.0 days (CI, 9.5–10.6 days) in mice receiving PBS, 10.2 days (CI, 9.8–10.7 days) in mice receiving Annexin V, 12.5 days (CI, 11.9–13.1 days) in mice receiving TRAIL, or 10.4 days (CI, 10.0–10.9 days) in mice receiving Annexin V + TRAIL, to 14.5 days (CI, 13.9–15.2 days) in mice receiving TP8 (12.5 mg/kg) (Figure 6D and Table 2). TP8 also caused a markedly delay of 6.9 days (CI, 5.5–8.4 days) in mice bearing A549 xenograft, 18.3 days (CI,15.7–21.0 days) in mice bearing Colo-205 xenograft and 10.0 days (CI, 7.5–12.5 days) in mice bearing Bel7402 xenograft for tumor growth to 1000 mm3 compared with PBS control (Figure 5C, Figure 6B, 6D and Table 2).

Table 2 Regression analysis for treatment effects on tumor growth
Figure 5
figure 5

Evaluation of antitumor effects of TP8 in nude mice bearing A549 carcinomas.

(A) Tumor volumes comparison among different groups (Mean ± SEM, n = 8, **p < 0.01 compared with each corresponding treatment). (B) Tumor doubling time comparison among different groups (Mean ± SD, n = 8, **p < 0.01 compared with each corresponding treatment). (C) Tumor growth delay time comparison among different groups (Mean ± SD, n = 8, **p < 0.01 compared with each corresponding treatment). (D) TUNEL analysis for tumor cells undergoing apoptosis within tumor tissues (original magnification × 400). The TUNEL-positive cells are indicated with red arrows.

Figure 6
figure 6

Evaluation of antitumor effects of TP8 on Bel7402 carcinoma and Colo-205 carcinoma in vivo.

Tumor volumes comparison among different groups in nude mice bearing (A) Bel7402 carcinomas and (C) Colo-205 carcinomas (Mean ± SEM, n = 8, *p < 0.05, **p < 0.01 compared with each corresponding treatment). Tumor doubling time and tumor growth delay time comparison among different groups in nude mice bearing (B) Bel7402 carcinomas and (D) Colo-205 carcinomas (Mean ± SD, n = 8, **p < 0.01 compared with each corresponding treatment).

TP8 induced tumor cell apoptosis in vivo

The tumor slices obtained from the mice in each group were subjected to TUNEL staining after the treatments were completed. The results showed an increase in the number of TUNEL-positive cells from the mice treated with TP8 compared with those from the mice treated with PBS, Annexin V, TRAIL or Annexin V + TRAIL (Figure 5D).

Discussion

The American Cancer Society reported that cancer is a widespread disease. In the US, approximately 1,660,290 new cases and 580,350 deaths are expected in 2013. Cancer is the second leading cause of death (next to heart disease) and accounts for 1 out of 4 deaths. Chemotherapy and radiotherapy are the two main approaches for cancer treatment besides surgery. P53 dysfunction, resistance caused by repeatedly administration of chemical agents or the severe side effects often result in the failure of tumor treatments.

TRAIL is a tumor-specific agent for cancer therapy, which is relatively safe when administered systemically compared with TNFα and FasL. Besides induction apoptosis of a variety of tumor cells, TRAIL elicits antitumor effects in xenograft models, including breast cancer, colon cancer, multiple myeloma, glioma and prostate cancer40,41. Nevertheless, the resistance to TRAIL-mediated apoptosis is a great challenge when it is applied in cancer treatment, because substantial numbers of cancer cells are resistant to TRAIL-induced apoptosis, particularly some highly malignant tumors. Combinations TRAIL with chemotherapy or radiotherapy are the main strategies used in clinic to overcome TRAIL-resistance. The modifications of TRAIL mainly focused on enhancement of tumor targeting of TRAIL. These approaches enhanced antitumor effects of TRAIL through increased concentration of TRAIL in tumor and did not increase TRAIL-sensitivity fundamentally. In this study, a completely new protein TP8 was developed to overcome TRAIL-resistance. The antitumor efficacy was greatly enhanced compared with TRAIL both in vitro and in vivo. In vitro, the EC50 of TP8 on a range of tumor cells, including TRAIL-sensitive cells, moderate-sensitive cells and TRAIL-resistant cells, was no more than 6 nM. To assess the in vivo antitumor activity of TP8, Colo-205, Bel7402 and A549 cells were selected to construct xenograft models, which also represent TRAIL-sensitive, TRAIL moderate-sensitive and TRAIL-resistant tumor cells, respectively. And the results revealed that TP8 effectively inhibited tumor growth and prolonged tumor doubling time and tumor growth delay time. These results above further indicated that TP8 greatly broadened antitumor spectrum of TRAIL both in vitro and in vivo.

The half-life of rhTRAIL is only 23 to 31 min in non-human primates and mouse tumor models, so high dose of rhTRAIL is needed to achieve antitumor effects due to its fast clearance42. And repeatedly administration of rhTRAIL at high dose may cause unexpected side effects. The recombinant replication-deficient adenoviral vector encoding human TRAIL (Ad5-TRAIL) was constructed to achieve better antitumor activity43. Contrary to rhTRAIL, TRAIL is expressed continuously after intratumor injection of Ad5-TRAIL and it's not necessary to be administered repeatedly. However, intratumor injection is not a convenient and conventional tumor therapy approach and intratumor-injected Ad5-TRAIL is unlikely to travel to other parts of the body, so it's less effective to eliminate metastasis. Agonistic anti-TRAIL-receptor antibodies, with the half-life of approximately 14–21 days, were developed to overcome the resistance caused by expression of TRAIL-R3/TRAIL-R440. Nevertheless, each antibody is specific to one death receptor and the expression of death receptors varies among tumor cells. Therefore, individualized detection of expressed death receptor impedes effective treatment with agonistic anti-TRAIL receptor antibodies. Compared with TRAIL, TP8 induced substantial apoptosis rapidly after administration, indicating that less TP8 protein was needed to achieve the same antitumor effects. Compared with Ad5-TRAIL, TP8 could travel to other parts of the body and eliminate metastasis effectively. Compared with agonistic anti-TRAIL receptor antibodies, TP8 could also induce apoptosis of tumor cells those expressed either DR4 or DR5. Additionally, TP8 could effectively prevent cytotoxicity caused by antibody crosslinks.

TP8 induced cell death in the same manner as TRAIL and both pathways were caspase-dependent. Through the increased activation of caspase-8, TP8 induced apoptosis of TRAIL-resistant A549 cells. Cys230 was critical to maintain the antitumor activity of both TRAIL and TP8. That's to say, Annexin V endowed the fusion protein TP8 with enhanced antitumor activity, without causing additional cytotoxicity.

In summary, a novel TRAIL chimeric protein TP8 was developed and both the in vitro and in vivo investigations showed promising results. All of our findings suggested that this novel approach could be used to overcome TRAIL-resistance and may have profound implications in cancer therapy.

Methods

Cell culture and reagents

A549, Bel7402, MDA-MB-231, MCF-7, HeLa, HepG2, U937, H446, H1229, PLC, Colo-205 and 293 T cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA). U937 and Colo-205 were cultured in RPMI 1640 (HyClone, Logan, UT). Others were cultured in DMEM (HyClone, Logan, UT). All the medium were supplemented with 10% FBS (HyClone, Logan, UT), 100 IU/ml penicillin-streptomycin (Invitrogen, Burlington, Canada). All the cells were grown at 37°C in 5% CO2.

EGFP-Annexin V, TRAIL, RGD-TRAIL and Annexin V were prepared in our lab.

Preparation of TP8 and TP8-V

The human TRAIL cDNA fragment (amino acid residues 114–281) was amplified by PCR from the pET32a-TRAIL plasmid and Annexin V cDNA fragment was amplified from pET28a-Annexin V plasmid, respectively. And the fragments were linked by overlapping PCR and cloned into pET28a expression plasmid (Novagen). The plasmids pET32a-TRAIL and pET28a-Annexin V were stored in our lab. Recombinant TP8 was expressed in Escherichia coli BL21 (DE3) cells (Novagen) and purified by metal affinity chromatography on Zn-sepharose and anion exchange chromatography on DEAE-Sepharose. All of the solutions were prepared with sterile, endotoxin-free water. Protein concentrations were determined using BCA Protein Assay Reagent (Galen Biopharm). After purification, TP8 was analyzed by SDS-PAGE under reducing (12%) and non-reducing (8%) conditions using Coomassie blue staining, while TRAIL was analyzed by non-reducing SDS-PAGE (12%). Cys230 is the activity center of TRAIL. In order to study the pathway TP8 induced apoptosis of tumor cells, we mutated the Cys230 of TRAIL domain of TP8 to Gly and this mutated protein was called TP8-V. TP8-V cDNA fragment was obtained from pET28a-TP8 plasmid by overlapping PCR and then cloned into pET28a expression plasmid. And TP8-V was purified in the same manner as TP8.

Cell viability assay

103–104 cells were seeded triplicate in 96-well plates. After 12 hours incubation, the Colo-205 and MDA-MB-231 cells, which are sensitive to TRAIL-induced apoptosis, were treated with various concentrations of TRAIL or TP8 for 4 hours. The A549, MCF-7, PLC, HepG2, Hela, U937, H446 and H1229 cells, which are moderate-sensitive or resistant to TRAIL-induced apoptosis, were treated for 24 hours. Cell viability was determined by MTT assay as described previously44. In brief, after drug treatment, MTT (5 mg/ml, 100 μl/well) was added to each well and after additional 4 hours incubation, the supernatant was removed and 100 μl/well DMSO was added. After shaking the plates for 10 min to dissolve the formazan, absorbance of each well was recorded with a microplate reader (Safire, TECAN, Switzerland) at 570 nm.

Clonogenic survival assay

A549 cells were seeded as single cells (300 cells/well) in 6-well plates, incubated 24 hours to allow adhesion and treated with same concentration (1.79 nM) of Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 24 hours. Cells were then washed twice and cultured in drug-free regular medium for additional 14 days. The medium was changed every five days. After the treatments were completed, the colonies formed were fixed with methanol and stained with Gimsa reagent. Cells treated with PBS were used as negative control.

Transmission electron microscopy analysis of cell morphology

Images of cell morphology were obtained using transmission electron microscope (H-60, HITACHI, Japan) as described before45. Briefly, the A549 cells were harvested and washed twice with PBS after these cells were treated with indicated concentrations (10.74 nM) of Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours. Afterward, the cells were incubated with 2.5% glutaraldehyde at 4°C overnight, dehydrated gradually with 70%, 80%, 90% and 100% ice-cold alcohol and embedded in EPON 812. The blocks were cut into ultrathin sections by ultramicrotome and stained with uranyl acetate/lead citrate. The cells treated with PBS were used as the negative control.

Apoptosis assay

Cells, treated with indicated concentrations of different proteins for indicated time, were stained with EGFP-Annexin V and propidium iodide (Sunshine Biotechnology, Nanjing, China) to measure the rate of apoptosis. Then the cells were analyzed using flow cytometer and the rate of apoptosis was analyzed with CELLQuest software (Becton Dickinson, Mountain View, CA). All of the experiments were three independent experiments and each experiment was repeated thrice. 293 T cells were used to assess cytotoxicity of TP8 against normal cells.

DRs binding ability of Annexin V, TRAIL, Annexin V + TRAIL and TP8 assay in vitro

Firstly, the proteins were labeled with FAM (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Then the A549 cells were incubated with equal molar concentrations of FAM-conjugated proteins at 37°C for 30 min, harvested, washed thrice with PBS and analyzed using flow cytometer. The fluorescence intensity was analyzed with CELLQuest software.

Western blot assay of signaling pathway

In order to study the changes of important proteins of apoptosis pathway, the A549 cells were treated with 10.74 nM Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours. After drug treatment, the cells were harvested and the whole cell lysates were prepared as described before46 with protease inhibitor cocktail (Roche, Milan, Italy). A total of 50 μg of protein was electrophoresed on 12% SDS-PAGE gels using a Tris-glycine system and transferred to polyvinylidene difluoride membranes (Millipore, Bedord, MA). The membranes were blocked with 5% non-fat dry milk in PBST for 1 hour and probed with antibodies against DR4 (IMGENEX, San Diego, CA), DR5 (IMGENEX, San Diego, CA), caspase-3 (Cell Signaling Technology, Beverly, MA), caspase-8 (Cell Signaling Technology, Beverly, MA), caspase-9 (Cell Signaling Technology, Beverly, MA), PARP (Oncogene) and GAPDH (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) overnight at 4°C. The blots were labeled with a secondary antibody (HRP-conjugated anti-rabbit or anti-mouse IgG) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and visualized with enhanced chemiluminescence detection reagents (Cell Signaling Technology, Beverly, MA).

Flow cytometer analysis of DR4 and DR5 expression levels

After treatment with PBS, Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours, the A549 cells were stained with fluorescence probe-conjugated antibodies: PE-anti-DR4 (E14958-101, eBioscience) or FITC-anti-DR5 (E11072-102, eBioscience). The expression levels of DR4 and DR5 were analyzed using flow cytometer. The fluorescence intensity was assessed with CELLQuest software. The cell staining experiments were repeated at least thrice.

Enzymatic activity assay of caspase-8 and caspase-3

A549 cells, treated with indicated concentrations (10.74 nM) of Annexin V, TRAIL, Annexin V + TRAIL or TP8 for 6 hours, were lysed with lysis buffer (Beyotime, Nantong, China). Caspase-8 activity was determined with caspase-8 assay kit (Beyotime, Nantong, China), in which Ac-IETD-pNA was used as substrate. And caspase-3 activity was determined with caspase-3 assay kit (Beyotime, Nantong, China), in which Ac-DEVD-pNA was used as substrate. For these caspase activity assays, the cell lysates were incubated with reaction buffer and substrates at 37°C for 1–2 hours and the absorbance was read at the wavelength of 405 nm. Caspase activity was evaluated by enzyme unit.

Xenograft experiments

Female BALB/c nude mice (5–6 weeks old), which were obtained from Beijing Animal Centre, were housed in environmentally controlled conditions (22°C; a 12 hours/12 hours light/dark cycle with the light cycle from 6:00 to 18:00 and the dark cycle from 18:00 to 6:00) and provided with pathogen-free food and water. Cancer cells were grown in medium as described before47 and 1 × 106 cells in 100 μl PBS were injected into the mid-right flank of mice, which developed tumors in 7–10 days with the size of approximately 50–100 mm3. For every experiment, the mice were randomized into 7 groups (8 mice per group) and intraperitoneally injected with PBS, 7.5 mg/kg Annexin V, 5 mg/kg TRAIL, 7.5 mg/kg Annexin V+5 mg/kg TRAIL, 6.25 mg/kg TP8, 12.5 mg/kg TP8 or 25 mg/kg TP8 daily for continuous 14 days. Tumor volume was measured every two days after the initial injection with calipers and determined as mm3 using the equation: A × B2 × 0.5247, where A was the length (mm) and B was the width (mm). Tumor doubling time refers to the time for a tumor to double in volume and tumor growth delay time is the time interval to reach 1000 mm3 compared with the PBS control group48.

Animal care and use were performed strictly in accordance with the ethical guidelines by the Nanjing University Animal Care and Use Committee and the study protocol was approved by the local institution review board.

TUNEL staining for tumor tissue sections

Tumor tissue sections (5 μm in thickness) were prepared according to standard protocols and then assessed for apoptosis. DNA fragmentation [terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) staining] was determined according to the manufacturer's instructions (Calbiochem, San Diego, CA).

Statistical analysis

Data are reported as mean with their 95% confidence intervals in parentheses. Statistical analysis was carried out using the Student's t-test. Data of apoptosis and caspase activity were expressed as mean ± SD and data of tumor sizes were expressed as the mean ± SEM. All the data were analyzed by use of SPSS software. The significance was shown as follows: *, P < 0.05; **, P < 0.01.