Introduction

Currently, gene therapy is a realistic option for all types of cancer treatment. Considering the poor survival rates due to drug resistance and lack of tumor specificity by using conventional methods, gene therapy offers a prospective approach for delivering the transgene to a desired location in human body. The chemical and biological gene delivery techniques are currently considered as model tools for cancer gene therapy in vivo. These approaches are less costly and more effective (higher gene transformation and no DNA degradation from serum nucleases) compared with mechanical and physical methodologies. Chemical methods are usually based on coating or incorporating a piece of nucleic acid (small interfering RNA or microRNA) on nanomaterials’ surface such as cationic polymers, cationic peptides and cationic liposomes,1, 2 whereas biological methods rely on viruses or bacteria as main delivery vectors. When using viral vectors, the key advantages are the super efficiency in invading and delivering therapeutic genes into the cell. Nevertheless, from a safety point of view, numerous viruses may cause insertional mutagenesis and systemic inflammation responses, thereby limiting clinical efficacy and restricting the range of applicable therapeutic approaches.3

The use of bacteria as gene delivery method for cancer therapy has been known for decades. The correlation between the tumor regression and bacterial infection dates back to nineteenth century.4 To date, different bacteria including both pathogenic strains and non-pathogenic strains have been tested in preclinical works.5 In short, it has been shown that bacteria organisms have the ability to replicate and target the hypoxic region of the tumor in vivo; usually only 3–5% of tumor cells are considered in the growth fraction, whereas the remaining 95% of tumor tissue is hypoxic to some degree.6 Nevertheless, there are notable differences in using pathogenic and non-pathogenic strains. For example, pathogenic bacteria such as Salmonella, Clostridium and Listeria have the capability of invading and amplifying gene encoding factors specifically within tumors.7, 8, 9 However, due to the safety reasons, this approach does not provide neither optimal nor prospective clinical approach. On the other hand, the use of beneficial bacteria such as Bifidobacterium strains offers a straightforward, safe and clinically acceptable approach for delivering therapeutic proteins locally within the tumor environment, external to tumor cells. Over the last decade, numbers of data have been published on the use of commensal gut bacteria that is, Bifidobacterium species such as breve, infantis, longum and adolescentis for the liver, lung and melanoma tumor therapy in vivo.10, 11, 12, 13

In the present work we have established a new recombinant B. breve strain expressing a human MDA-7/IL-24 gene for induction of head and neck squamous cell carcinoma (HNSCC) apoptosis in vivo. MDA-7/IL-24 is a member of the interleukin (IL)-10 family and has been strongly correlated with apoptosis induction in a variety of cancer including melanoma, breast, liver, prostate, ovarian and nasopharyngeal cancer.14, 15, 16, 17, 18, 19 In our previous study we have demonstrated that IL-24 could be a good prognostic biomarker and a valuable indicator of second primary malignancies in HNSCC.20 According to our knowledge, this will be the first time of using Bifidobacterium species as delivery strategy of IL-24 for cancer therapy in vivo.

Results

Construction of B. breve recombinant strains

The recombinant Escherichia coli–Bifidobacterium shuttle vectors pLW5 and pLW9, which include IL24 (835 bp) and green fluorescent protein (GFP, 930bp) fragments, respectively, were used to obtain B. breve-IL24 (pLW5-breve) and B. breve-GFP (pLW9-breve) transformants. Contrary, recombinant strain B. breve-Control was transformed using the pAM4 (naked plasmid). The correct construction patterns were analyzed by digesting plasmids isolated from B. breve transformants with Xbal and Xhol (Figure 1a). Differences in bacteria growth rate between the wild-type strain and the transformants were analyzed. Briefly, the results showed higher growth rate in B. breve-IL24 and B. breve-GFP compared with B. breve-Control, but no difference in comparison to B. breve wild-type strain, during the exponential phase. Furthermore, during the last growth stage, there was no difference among the four observed groups (Figure 1b). In addition, as shown in Figure 1c, no difference was observed in bacteria shape between transformants and the wild-type bacteria.

Figure 1
figure 1

New recombinant B. breve transformants construction, growth curve and Gram staining difference compared with B. breve wild-type strain. (a) Enzyme (Xbal/Xhol) digestion pattern of plasmids isolated from B. breve-Control (L1), B. breve-IL24 (L2) and B. breve-GFP (L3). For Breve-IL24 and Breve-GFP, 835 bp and 930 bp fragment were obtained individually as designed. (b) Growth curve of B. breve wild type compared with the transformants over 72 h. (c) A gram staining of B. breve wild type and its transformants, magnified by × 100 oil. B. breve 1.3001 and its transformants were often branched, Gram-positive bacterium.

Full size image

Target gene expression in vitro

Real-time PCR was used to investigate the IL-24 mRNA expression level in the B. breve-IL24 and B. breve-GFP strains, respectively; all data were compared with the B. breve-Control (naked plasmid) strain. Briefly, the results showed that the mRNA level of IL24 was 9466.72±440.241-fold in B. breve-IL24 group compared with 2.76±1.059-fold registered in B. breve-GFP and 1.06±0.502-fold registered in the B. breve-Control strain (Figure 2a). The P-value and multiple comparisons were shown in Figure 2b.

Figure 2
figure 2

IL-24 gene expression between B. breve-Control and other two transformants (B. breve-IL24 and B. breve-GFP). *P<0.05. (a) Fold changes of IL-24 expression in mRNA level among B. breve-Ctrl, B. breve-GFP and B. breve-IL24. (b) The P value and multiple comparisons among three groups.

Full size image

Tumor-targeting property of recombinant B. breve in vivo

Tumor-specific targeting of the new recombinant B. breve strains were monitored in vivo, over 1, 3 and 7 days post injection (Figure 3). Briefly, 24 h post injection, bacteria clones were detected in the heart, liver, spleen, lung, kidney and tumor of the nude mice. Over the next 7 days, growing clones were visible only in the liver and tumor (day 3) and afterwards only in the tumor tissue (day 7). These results suggested that our new recombinant B. breve strain has the capability of targeting the tumor tissue in vivo.

Figure 3
figure 3

Analysis of targeting property of B. breve transformants to tumor tissue in vivo. (a) Day 1: bacteria visible in all the main organs including tumor tissue, × 10 dilution. Day 3: bacteria visible only in the liver and tumor tissue, × 10 dilution. Day 7: bacteria visible only in the tumor tissue, × 10 dilution. (b) The number of bacteria clones in different organs over Day 1, 3 and 7.

Full size image

B. breve-IL24 induce tumor apoptosis and tumor growth inhibition in vivo

Tumor size was measured to monitor the growth inhibition efficiency in vivo. Saline group was used as blank control and B. breve-GFP was used as negative control. Our data suggested a higher tumor growth inhibition in B. breve-IL24 compared with B. breve-GFP group and higher tumor growth inhibition in B. breve-GFP compared with the Saline group, Figure 4a. The P-value among each group was shown in Figure 4b. In general, there were statistical differences in tumor size between Saline group and B. breve-GFP group from day 9. The statistical differences between Saline group and B. breve-IL24 group in tumor size was from day 3, whereas no significance between B. breve-GFP and B. breve-IL24 group were observed. Nevertheless, the average tumor size (including slower tumor growth rate) in B. breve-IL24 was smaller compared with the B. breve-GFP group.

Figure 4
figure 4

Growth curve and apoptosis induction in mice bearing a WSU-HN6 tumor, after treatment with B. breve-IL24, B. breve-GFP and the Saline in vivo. (a) Tumor growth curve over 2 weeks of therapy with different therapeutic approach (n=10). (b) P-value among each group. (c) Tumor uptake of [18F]-ML-10 tracer in vivo, by PET/CT scanning; the detected radioactivity was 135.35 for B. breve-IL24 group (volume 169.13 mm3) and 20.09 for the Saline group (volume 129.96 mm3). (d) Anti- and Pro- apoptotic protein expression post therapy, in ex vivo.

Full size image

When used [18F]-ML-10, a Positron Emission Tomography (PET) tracer for apoptosis, our micro Positron Emission Tomography/Computed Tomography (PET/CT) scan showed that a tumor treated with B. breve-IL24 has significantly higher apoptotic induction (namely higher radioactivity) compared with the Saline group in vivo, Figure 4c. To explore the underlying mechanism of apoptosis, anti- and pro- apoptotic protein expression was tested ex vivo, by western blot approach. In general, pro-apoptotic protein Bim was highly expressed in B. breve-IL24, whereas little Bcl-2 expression was detected in this strain compared with B. breve-GFP and Saline groups, Figure 4d. In addition, the expression of caspase 3 and cleaved caspase-3 was elevated in the B. breve-IL24 group. Nevertheless, the expression of cleaved caspase-3, which is the active component of caspase-3, was difficult to test when using the total caspase-3 antibody analysis (Supplementary Figure S1 in Supplementary Information). To conclude, this data suggested that the IL-24 expressing new recombinant B. breve-IL24 strain may be a suitable approach for cancer therapy in vivo.

In addition, compared with Saline group, higher Bim and lower Bcl-2 expression (but no cleaved caspase-3) was observed in B. breve-GFP group. This result was consistent with the tumor growth inhibition rate, thus suggesting the ability of bacteria itself to induce tumor inhibition in vivo.

Discussion

Mda-7/IL-24, a member of the IL-10 family, has been shown to function as a cytokine at physiological level and to exhibit anti-cancer effects at supra-physiological level. During physiological conditions, IL-24 protein can interact in a paracrine manner with IL-22R1/IL-20R2 and IL-20R1/IL-20R2 receptors expressed in immune system cells, resulting in the release of secondary cytokines.21 These cytokines might then produce additional changes in cellular physiology, including cell proliferation in specific immune cell subsets, antitumor immune responses, and/or effects on immune cell growth. At supra-physiological level, number of studies have identified IL-24 as a multi-functional protein, which affects a biological behavior of cancer.14, 15, 16, 17, 18, 19 These functions include apoptosis and/or autophagy induction, tumor angiogenesis inhibition, migration and invasion inhibition and bystander effect, whereas no effect was detected in healthy cells. According to previous studies, Phase I clinical trial by Ad.mda-7 (INGN 241), which employed intratumoral injection, was initiated to evaluate its safety profile and its biological effects, both locally and systemically in humans with multiple advanced cancers.21 According to Fisher and colleagues,22 future clinical trials with IL-24 should employ improved delivery vectors and should focus on systemically targeted delivery.

From our previous study, we have learned that IL-24 expression can be strongly associated with the survival rate and second primary malignancies occurrence in HNSCC.20 In addition, we have showed that apoptotic protein expression in HNSCC cell could be induced by adenoviral delivery of IL-24 (Ad-IL24) in vitro.20 Nevertheless, it was indispensable to continue working on discovering the safer and more efficient gene delivery methods due to safety reasons (possible cause of insertional mutagenesis and systemic inflammatory). In this study, we have developed a novel delivery system for antitumor gene IL-24 by using probiotic bacteria B. breve for tumor therapy in vivo.

Restriction and modify system was a key factor when constructing the shuttle vector between E. coli and Bifidobacterium strains.23 hup promoter and sec2 fragment were chosen from B. breve UCC2003 strain, in order to facilitate the target gene IL-24 replication and secretion outside the cells. The addition of raffinose to the culture medium, tested as a good carbon source for B. breve, assisted to enhance the B. breve cell growth in vitro.24, 25 After construction of the transformants, no significant differences in cell growth and cell shapes were found between the wild type and recombinant strains. In addition, enzyme digestion of the plasmids extraction from the recombinant bacteria was completely in line with design.

In vitro expression of target gene tests showed that IL-24 were highly expressed in B. breve-IL24 at mRNA level. In addition, the GFP protein expression from B. breve-GFP was tested in vivo. After subcutaneous injection, GFP signal was clearly visible as detected by whole-body fluorescence reflectance imaging device (Supplementary Figure S2 in Supplementary Information). To sum up, these data have demonstrated the ability of our new recombinant strains, B. breve-IL24 and B. breve-GFP, to constitutively express IL-24 or GFP genes under the direct action of the hup promoter.

Moreover, the biodistributions of our recombinant B. breve-GFP and B. breve-IL24 were further investigated in vivo. Seven days post bacteria injection, bacteria was detected only in the tumor tissue. These results have re-confirmed the recently published data by proving the specific capability of the Bifidobacterium strain to target and colonize the tumor tissue in vivo.26 Furthermore, Cronin et al.27 have recently discovered that Bifidobacterium may translocate from the gastrointestinal track, enter the bloodstream and target the tumor tissue in mice. However, the route and mechanism of this translocation still remain unclear. The initial research design of our study was to use the GFP reporter gene to investigate the biological behavior of B. breve-GFP in vivo and in real-time. Unfortunately, due to low wavelength (causing higher tissue scattering and autoflorescence in vivo), we were unable to detect the GFP signal in deep tissue. Recently, we have tested and monitored the biologic behavior of bacteria expressing a novel far red fluorescence reporter gene mKate2 in the gut of the mice.28 Owing to lower autoflorescence and lower tissue scattering (above 600nm) in deep tissue, far red fluorescence is a suitable choice for our further examination of the route and mechanism of B. breve translocation and tumor targeting in the small animals such as mice.

Finally and most importantly, our new recombinant strain B. breve-IL24 has shown to be an promising therapeutic approach for cancer, by promoting tumor apoptosis which in turn promotes inhibition of tumor growth in vivo. As expected, higher tumor growth inhibition rate was observed in the Drug group (treated with B. breve-IL24) compared with the Saline and B. breve-GFP group; even though, no statistical significance was observed in tumor size between B. breve-GFP and B. breve-IL24 group. As illustrated in Results section, the average tumor size (including slower tumor growth rate) in B. breve-IL24 was smaller compared with B. breve-GFP group. We also analyzed the bacterial therapy on an additional tumor cell line (that is, melanoma cell line MDA-MB-435s). The tumor growth curve and P-value are provided in the Supplementary Data and Supplementary Figure S3 in Supplementary Information. As the results in HNSCC, the average tumor size (including slower tumor growth rate) in B. breve-IL24 was smaller compared with B. breve-GFP group in melanoma. In addition, the apoptotic induction of B. breve-IL24 was examined in real-time, by using [18F]-ML-10 tracer and a homemade PET/CT in vivo imaging device. [18F]-ML-10 is a PET imaging probe specifically designed to target and enter apoptotic cells in vivo, and only recently it was used for the first time in a human study.29 In this study, we used [18F]-ML-10 to compare a tracer/tumor uptake in the Saline and Drug group mice bearing a similar tumor size. Our imaging data showed higher tracer uptake in the Drug group compared with the Saline group. Although the data have been obtained from a smaller number of mice, these preliminary results suggested strong apoptotic induction by B. breve-IL24 therapy in vivo. Moreover, PET/[18F]-ML-10 may be a suitable imaging approach for further studying the apoptotic induction by B. breve-IL24 during early stages (few days post injection) of tumor development in vivo.

In addition, further ex vivo analysis of tumor tissue has confirmed that B. breve-IL24 could reduce the anti-apoptotic protein Bcl-2 and enhance pro-apoptotic protein Bim expression. Moreover, compared with Saline group, lower Bcl-2 and higher Bim expression were detected in tumor treated with B. breve-GFP negative control strain. This data indicated that the B. breve itself have the ability to influence the tumor microenvironment and activate the host immune response in vivo, which was already verified by other researchers. Sivan et al.30 have demonstrated that oral administration of commensal Bifidobacterium alone could enhance antitumor immunity in vivo to the same degree as programmed cell death protein 1 ligand 1-specific antibody therapy. The underlying mechanisms could act trough altering dendritic cell activity.30

Moreover, elevated expression of total caspase 3 was observed in B. breve-IL24 compared with other two groups. In addition, a shadow dye in a cleaved caspase-3 position (19 and 17 kDa) was observed on the same polyvinylidene difluoride membrane. Thus, we hypothesized that the high signal of pro-caspase-3 was shedding the weak signal of cleaved caspase-3. Therefore, we analyzed the expression of cleaved caspase-3 independently. Our data indicated a high expression of cleaved caspase-3 in B. breve-IL24 group (Supplementary Figure S1 in Supplementary Information). To conclude, considering the weaker expression of cleaved caspase-3 compared with obvious higher expression of Bim and lower expression of Bcl-2 in the B. breve-IL24 group, we concluded that the apoptosis effect by this strain may mainly rely on the mitochondrial control (Bcl-2 and Bim) of caspase-independent cell death.31

Currently, several engineering bacteria have been under clinical trial investigation. For example, BioMed Valley Discoveries, Inc. (Kansas City, MO, USA) is currently examining the safety and the anti-tumor activity of intratumoral administration of Clostridium Novyi-NT spores in patients with treatment-refractory solid tumor malignancies (https://clinicaltrials.gov/ct2/show/NCT01924689). In addition, Marina Biotech, Inc. (City of Industry, CA, USA), has recently developed a new drug CEQ508 (E. coli bacteria engineered using transkingdom RNA interference platform), which can be administrated orally, and is used for Familial Adenomatous Polyposis treatment in clinical practice (http://www.marinabio.com/files/5014/3880/5803/15-03-30_-_Marina_Biotech_Announces_FDA_Fast_Track_Designation_for_CEQ508.pdf). Recently, Cronin et al.27 compared different administration routes of B. breve UCC2003, that is, oral intake vs intravessal injection using a murine animal model. Their data suggested that orally taken of B. breve UCC2003 was similar with intravenous administration in subsequent homing to and growth specifically in tumors.27 Moreover, we are currently investigating the oral intake of B. breve-IL24 in vivo.

To conclude, a new recombinant strain B. breve-IL24 can induce apoptosis and consequently increase tumor growth inhibition in HNSCC mice model in vivo. In addition, B. breve-IL24 is not affecting the healthy tissue, thus could offer a safe and clinically acceptable therapeutic approach for tumor therapy in vivo.

Materials and methods

Reagents

The HNSCC cell line (WSU-HN6), authenticated by short tandem repeat profiling (Supplementary Figure S4) and free of mycoplasma contamination, was obtained from Central Laboratory of Peking University School and Hospital of Stomatology. It was cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Invitrogen, Waltham, MA, USA) and 1% Penicillin/Streptomycin (Gibco, Waltham, MA, USA) in a humidified atmosphere containing 5% CO2/95% air at 37 ºC. B. breve (No. 1.3001) was purchased from China General Microbiological Culture Collection Center (Beijing, China) and routinely cultured for 72 h in MRS Broth (BD, Franklin Lakes, NJ, USA) or RCM (Oxoid, Waltham, MA, USA) plus Raffinose (Amresco, Solon, OH, USA) supplement and 0.5% L-cysteine▪HCl (Sigma Aldrich, Damstadt, Germany) (MRSRC), under anaerobic environment at 37 °C. E. coli DH5α strain was cultivated in Luria-Bertan broth at 37 °C with vigorous shaking. Bacteria culture medium was supplemented with the following antibiotic concentration: 100 μg ml−1 Ampicillin, 5 μg ml−1 Tetracycline for E. coli and 5 μg ml−1 Tetracycline for Bifidobacterium strain. All the restriction enzymes and T4 DNA ligase were purchased from Thermo Fisher Scientific (Waltham, MA, USA).

Shuttle vector construction

B. breve UCC2003 was a kind gift from Prof. Douwe van Sinderen.7 A backbone plasmid pAM4 was a kind gift from Prof. Baltasar Mayo.32 Cloning vector pBluescript SK was stored in ACTA. Genomic DNA of B. breve UCC2003 was purified following the manufactures protocol (GeneJET Genomic DNA Purification Kit, Thermos Fisher), which was the template of hup promoter and sec2 fragment. IL-24 (Genebank NC_000001.11) fragment was PCR amplified using Human Universal cDNA (Clontech) as template. Briefly, a shuttle vector pLW5 was created as follows: promoter hup and signal peptide sec2 fragments were first sequentially cloned into pBluescript SK (pBSK) plasmid. IL24 gene was then cloned into pBSK-hup-sec2, by obtaining pBSK-hup-sec2-IL24 vector. Finally, the fragment hup-sec2-IL24 was digested and cloned into a backbone plasmid pAM4, which resulted in a shuttle vector pLW5. Additionally, pLW9 was constructed by cloning GFP into pLW5, which replaced sec2 and IL24 gene fragments.

Transformation of Bifidobacterium breve

An overnight B. breve (No.1.3001) culture was first re-inoculated into fresh MRSRC medium (1:10) in anaerobic environment at 37 °C until the OD600 of 0.4–0.6 was reached. The bacteria were then collected and centrifuged at 5000 rpm (4 ºC), washed 3 times using washing buffer (0.5M sucrose+1 mM citrate acid) and finally re-suspended in 1:250 diluted washing buffer, resulting in B. breve competent cells.

0.5 μg plasmid (pAM4 or pLW5 or pLW9) was mixed with 50 μl competent cells in a pre-cooled Gene Pulser cuvette (Bio-rad, 1mm) and chilled on ice for 30 min before electroporation. A Gene Pulser Apparatus was employed and used at a high voltage of 15 KV/cm, 25 μF capacity and a parallel capacity of 200 Ω. 950 μl MRSRC medium was added to the cuvette immediately post electroporation and cultured for another 3 h at 37 °C in anaerobic environment to recover the antibiotic expression. The bacteria were then plated on the RCM agar plates supplement with 5 μg ml−1 Tetracycline and cultured at 37 °C in anaerobic environment for 48–72 h.

Culture of transfected Bifidobacterium breve and detection of positive clones

Transformed bacteria (B. breve-Control, B. breve-GFP, B. breve-IL24) were collected and cultivated in MRSRC medium supplement with 5 μg ml−1 Tetracycline for 48–72 h, afterwards the plasmid DNA was extracted (Tiangen, Beijing, China). Xbal and Xhol enzyme digestion was employed to test the extracted plasmid. Gram staining and growth curve were used to examine the bacteria shape and growth rate change post transformation.

Real-time PCR

To investigate the expression of IL24 in new recombinant Bifidobacterium strains, total RNA was extracted by using RNAprep Pure Cell/Bacteria Kit (Tiangen, Beijing, China), which was reversely transcribed into double-stranded cDNA using M-MLV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer’s protocols. Real-time PCR using SYBR green reagent (Roche, Indianapolis, IN, USA) was performed in an ABI Prism7000 Sequence Detection System (Applied Biosystems, Life Technologies, Warrington, UK). The primers for IL24 were as follows: IL24-F: TGCTGGAGTTCTACTTGAA, IL24-R: AGTTGTGACACGATGAGA. We choose uvrD as internal reference, which TTACCTATGATGCCTTCTTC as forward and CTGAGTGGCTGAGTATTC as reverse primers. The standard Real-time PCR conditions were 10 min at 95 °C, followed by 40 cycles of 95 °C for 15 s, and 60 °C for 1 min. All reactions were performed in triplicate. The expression level of the target transcript in each sample was calculated using the comparative 2−ΔΔCt method after normalization to uvrD expression. B. breve-Control, which was empty plasmid pAM4 transformants, was used as control group when comparing gene expression. The experiment was repeated three times.

Animal experiments

This study was approved by the Medical Ethical Committee of Peking University Health Science Center (No. LA2016276). Balb/c male nude mice, 6–8 weeks old, weighing 19–21 g, were obtained from Vital River Laboratories, China. All the animals were housed in an environment with the temperature of 22±1 ºC, relative humidity of 50±1% and a light/dark cycle of 12/12 h. All animal studies (including the mice euthanasia procedure) were done in compliance with the regulations and guidelines of Peking University Institutional Animal Care and Use Committee. When compared three groups’ tumor inhibition efficiency, the sample size was calculated according to one-way ANOVA (two-sided) model formulas provided in the website http://powerandsamplesize.com/Calculators/Compare-k-Means/1-Way-ANOVA-Pairwise. α=0.05 and β=0.9 was chosen. The sample size was limited to seven to ten after calculating to allow for the detection of highly significant differences.

Biodistribution study

12 mice were subcutaneously injected with 5 × 106 WSU-HN6 cells on the mice back (close to the right shoulder). When the tumor tissue reached approximately 5 mm in diameter, mice were randomly dividend into the B. breve-GFP and B. breve-IL24 group, 6 mice per group. 1 × 108 viable recombinant bacteria were then collected, washed 3 times, re-suspended in 200 μl phosphate buffered saline (PBS) and injected by tail vain. The biodistribution experiment was performed on day 1, 3 and 7 post bacteria injection. During each time-point, 2 mice per group were euthanized, the main organs were dissected, weighed, and homogenized after added 9 parts of PBS containing 0.1% Tween 20 per g of per organ (dilution 1:10), assuming that the volume of 1 g of organ corresponds to 1 ml of PBS. Next they were plated on RCM agar plus tetracycline plates in different dilution times (1:10, 1:100, 1:1000).

Cancer therapy in vivo

30 mice were subcutaneously injected with 1x106 WSU-HN6 cells on the mice back. When the tumor tissue reached approximately 50 mm3 in size, mice were randomly divided into three groups (Drug, GFP and Saline group), 10 mice per group. There was no significance difference of the average tumor sizes among three groups after random allocation. The therapy was given by tail vein twice per week, for a total of 2 weeks, as follows: 1 × 108 B. breve-IL24 in 200 μl for the Drug group; 1 × 108 B. breve-GFP in 200 μl for the GFP group; 200 μl PBS for the Saline group. Tumor size was measured by researcher blinded to the group allocation twice a week post bacteria injection, using standard caliper tumor volume calculation method Vtumor= (length x width2)/2. After treatments all mice were euthanized and tumor tissues were dissected and quick frozen in liquid nitrogen, and then stored at −80 °C for future western bolt assay. This experiment was run in triplicate.

PET/CT imaging

The tumor apoptosis were analyzed by our homemade small PET/CT device, using [18F]-ML-10, which is a novel PET tracer for apoptosis.33 Briefly, 0.3 mCi [18F]-ML-10 was first injected by tail vein (3 mice/group). 10 min later, mice was injected with Avertin anesthesia (2.5%, 0.5 ml per mice, i.p.) and palced on the animal bed in the dorsal position. During the next 54 min, the whole body PET/CT images were acquired as follows: 2 bed for CT scanning (15 min/ bed) and 3 bed for PET scanning (8 min/bed).

Western blot

Tumor tissues were carefully dissected and stored in liquid nitrogen, followed by lysing in RIPA buffer (Applygen, Beijing, China) containing proteinase inhibitors and phosphatase inhibitors. After measuring protein concentration using the BCA kit (Thermo Fisher Scientific, Waltham, MA, USA), equal amounts of protein samples were separated by 15 % SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes by wet blotting. The membranes were blocked in 5 % non-fat dry milk for 1 h and probed with antibodies against caspase-3 (8G10) (CST 9665, 1/1000), cleaved caspase-3(Asp175) (CST9661, 1/1000), Bim (GTX27888, 1/1000), Bcl-2 (CST3498, 1/1000) and GAPDH (CST8884, 1/1000) separately at 4 °C overnight. After incubation with peroxidase-linked secondary antibodies for 1 h at room temperature, the enhanced chemiluminescent reagent was used to visualize the immune-reactive proteins.

Statistical analysis

Real-time PCR results and tumor size were summarized as mean±s.d. Regarding IL-24 expression and tumor size, comparisons were made using one-way analysis of variance (two-side) after normal distribution test and homogeneity of variance analysis. Depending on sample variance, least significant difference was used for multiple comparisons and no adjustments were made in tumor size analysis and Games–Howell’s correction was applied for multiple comparisons in IL-24 expression level. All calculations and analyses were performed using SPSS 19.0 software (SPSS Inc., Chicago, IL, USA). When assessing, investigator was blind to the group allocation. A P-value<0.05 was considered statistically significant.