Anti‐tumour effects of a dual cancer‐specific oncolytic adenovirus on Breast Cancer Stem cells

Abstract Apoptin can specifically kill cancer cells but has no toxicity to normal cells. Human telomerase reverse transcriptase (hTERT) can act as a tumour‐specific promoter by triggering the expression of certain genes in tumour cells. This study aims to investigate the inhibitory effects and to explore the inhibitory pathway of a dual cancer‐specific recombinant adenovirus (Ad‐apoptin‐hTERTp‐E1a, Ad‐VT) on breast cancer stem cells. Breast cancer cell spheres were obtained from MCF‐7 cells through serum‐free suspension culture. The cell spheres were detected by flow cytometry for CD44+ CD24− cell subsets. The stemness of MCF‐7‐CSC cells was confirmed by in vivo tumorigenesis experiments. The inhibitory effect of the recombinant adenoviruses on MCF‐7‐CSC cells was evaluated by CCK‐8 assay. In addition, the stemness of adenovirus‐infected MCF‐7‐CSC cells was analysed by testing the presence of CD44+ CD24− cell subsets. The ability of the recombinant adenovirus to induce MCF‐7‐CSC cell apoptosis was detected by staining JC‐1, TMRM and Annexin V. Our results showed that a significantly higher proportion of the CD44+ CD24− cell subsets was present in MCF‐7‐CSC cells with a significantly increased expression of stem cell marker proteins. The MCF‐7‐CSC cells, whlist exhibited a strong tumorigenic ability with a certain degree of stemness in mice, were shown to be strongly inhibited by recombinant adenovirus Ad‐VT through cell apoptosis. In addition, Ad‐VT was shown to exert a killing effect on BCSCs. These results provide a new theoretical basis for the future treatment of breast cancer.


| INTRODUC TI ON
Cancer, a complex disease characterized by uncontrolled cell growth and metastasis, has caused millions of fatalities annually worldwide.
The development of tumours is contributed by different factors, either externally (unhealthy lifestyles and pollutants in the living environment), or internally (hormonal changes, immune response and genetics). 1 Both internal and external factors can contribute to the onset of cancer either individually, or collectively. Tumour development can also be accompanied by additional factors such as cancer stem cells (CSCs; also known as tumour/cancer initiation cells or stem cell-like cancer cells) that are responsible for initiating cancer. There are different tumour subtypes with different morphologies and behaviours, although the source of tumour heterogeneity remains unclear. CSCs are a specific subset of cancer cells with basic self-renewal and differentiation properties, which can produce tumours with genomic and phenotypic heterogeneity. 2 Accordingly, tumour heterogeneity and malignancy progression are contributed by a small number of cancer cells with stem cell-like properties, rather than by the vast majority of cancer cells. Firstly proposed in leukaemia, 3 CSCs have later been found to be present in solid tumour breast cancer. 4 These stem cell-like breast cancer cells display expression markers similar to those found in pluripotent progenitor cells, suggesting that these breast cancer stem cells (BCSCs) may be derived from mammary stem cells (MaSCs). 5 The presence of CSCs has been reported to occur in almost all types of malignant haematological tumours and solid tumours. 5 6 The migration behaviour of cancer stem cells is one of the biggest challenges cancer treatment, because migratory cells can not only change their morphology, they are resistant therapeutic drugs. Hence, one of the potential methods for the treatment of cancer is targeted therapy, which utilizes different specific recognition markers and signal transduction pathways, separately or combined with targeted quiescent phases, drug efflux cells and apoptosis-resistant cells. 7 Differences in the expression profile of cancer markers can be used to facilitate the design of treatment strategies.
CSCs have been shown to promote tumour initiation and metastasis, leading to therapeutic resistance. 8 Consequently, breast cancer with a high proportion of CSCs is associated with poor treatment outcomes. Some studies have shown that BCSCs exhibit relative resistance to conventional treatment in preclinical models and clinical trials. Consistently, at in vitro settings, paclitaxel was observed to be enriched for cells expressing BCSC phenotypes due to chemotherapy resistance. 9 Clinical studies have shown an increased expression of CD44 + CD24 − phenotype in primary tumours after chemotherapy. 10 It has also been found that BCSCs in different breast cancer cell line models are resistant to radiation therapy. 11 In addition, preclinical studies have demonstrated that an enhanced resistance to endocrine therapy is associated with the increased proportion of BCSCs, 12 indicating that BCSCs have an intrinsic resistance to anticancer therapy. 13 The inactivity nature of CSCs renders the cells to be insensitive to DNAdamaging agents and radiation. In breast cancer, it has been shown that CSCs contribute to malignant progression, suggesting that BCSC targeting may improve therapeutic efficacy. Firstly proposed in 2003, CD44 + CD24 − phenotype is by far the most commonly used marker for characterizing BCSCs. 4 Given that BCSCs contribute to a large number of primary tumours, the combination of BCSC-targeting agents and individualized breast cancer treatment can improve clinical outcomes.
Apoptin is a protein expressed by chicken anaemia virus that induces apoptosis in tumour cells. In normal cells, Apoptin is localized in the cytoplasm and does not cause apoptosis; whilst in tumour cells, Apoptin is localized in the nucleus. The nuclear localization of Apoptin is closely associated with the induction of tumour cell apoptosis. Apoptin is an attractive candidate for targeted tumour therapy, as it can specifically induce tumour cell apoptosis and is non-toxic to normal cells. Meanwhile, human telomerase (hTERT) is a ribonucleoprotein that plays a role in cell senescence and immortalization, so unlimited proliferation is observed in cells with highly expressed hTERT. The promoter of hTERT appears to be inactive in most normal cells, but it is highly active in most tumour cells. The high efficiency and specificity of hTERT promoter make hTERT an effective target in tumour cells for oncolytic adenovirus.
A dual cancer-specific oncolytic adenovirus Ad-Apoptin-hTERTp-E1A (Ad-VT) with the Apoptin and hTERT promoter was constructed together with three control viruses Ad-Apoptin (Ad-VP3), Ad-hTERTp-E1A (Ad-T) and Ad-Mock. 14 The constructed cancer-specific oncolytic adenovirus Ad-VT can specifically replicate in tumour cells to induce cell death, because apoptin is a tumour-specific killing protein, in addition, the hTERT promoter specifically triggers the adenovirus replication essential gene E1a. In this study, serum-free suspension culture was used to identify CD44 + CD24 − breast cancer stem cells with high tumorigenic ability through enhanced expres-

| Isolation and culture of breast cancer stem cells
MCF-7 cells were digested and centrifuged (300 ×g, 5 minutes) before suspension in serum-free medium (SFM, ie 20 ng/mL bFGF, 20 ng/mL EGF added to DME/F12, 2% B27, 1% N-2) followed by inoculation in ultra-low adhesion 6-well plates at a density of 2500 cells/mL for culturing at 37°C with 5% CO 2 . The process of cell pellet formation was observed, and the culture medium was changed at an appropriate time.
After culturing for 7-14 days, the cell pellets were collected by centrifugation before resuspension by Accutase and then digested at 37°C with 5% CO 2 for 10 minutes. After digestion, the cells were centrifuged at 300 ×g for 10 min before resuspension in SFM and followed by inoculation in ultra-low adhesion 6-well plates at a density of 2,500 cells/ml for culturing at 37°C with 5% CO 2 . 10 µL of CD44-FITC and 10 µL of CD24-PE. The cells were then incubated at 4°C for 30 minutes in the dark before centrifugation at 300 ×g for 10 minutes, followed by a cleaning step twice with 1 mL of fluorescent lotion, and a fixing step with 500 µL of fluorescent fixative (PBS containing 4% formaldehyde). The cells were suspended and then transferred to a flow tube for detection.

| Detection of the expression of tumour stem cells markers by Western blot
The marker proteins of tumour stem cells were detected by immunoblotting. MCF-7-CSC and MCF-7 cells were trypsinized and collected by centrifugation at 1875 ×g for 5 minutes. The cell pellets were resuspended in lysis buffer, and the protein solution was collected by centrifugation at 10800 ×g for 5 minutes. All samples were analysed by Western blot. 15

| In vivo tumour formation of MCF-7-CSC
NOD/SCID mice were randomly divided into 9 groups. The mice in each group were subcutaneously injected with 2 mg/mL of oestradiol benzoate at a weight ratio of 1 mg/kg, once every 5 days at the back 3 days prior to cell inoculation in order to establish a tumourbearing model. After disinfection using alcohol cotton balls, 100 µL Tumour formation was observed within 60 days of cell inoculation in the mice. The largest and the shortest tumour diameter, as well as the tumour volume, was calculated using the following formula: 0.52 14,16,17 The tumour growth trend in mice was analysed. The tumour volume doubling time was calculated using the formula: TVDT = t×[lgV t -lgV 0 )], where V 0 is the previous tumour volume, and V t is the next tumour volume and t is the time difference between two volume measurements.

| CCK-8 assay
MCF-7-CSC cell suspensions were adjusted to a density of 1 × 10 5 cells/mL before cell SAP was separately added to a 96-well ultralow adhesion culture plate at 100 μL/well (six replicate and control wells) and then cultured at 37°C with 5% CO 2 for 24 hours.
Then, three recombinant adenoviruses Ad-VT, Ad-T, Ad-VP3 and Ad-Mock were subsequently inoculated at 100 MOI, 10 MOI and 1 MOI, respectively. At 24, 48, 72 and 96 hours, the cells were added with 10 µL of CCK-8 in the dark, and then incubated for 1-4 hours at 37°C with 5% CO 2 before absorbance measurement at 450 nm using a microplate reader. Cell inhibition rate was calculated using the following formula: cell inhibition rate = [(Ac-As)/ (Ac-Ab)] × 100%, where As is the experimental well containing cells and recombinant adenovirus with CCK-8 added; Ac is the control well containing cells but no recombinant adenovirus with CCK-8 added; Ab is the blank well with only CCK-8 but without cells and recombinant adenovirus.

| Evaluation of the effect of recombinant adenovirus on the CD44 + CD24 − cell subsets of MCF-7-CSC by flow cytometry
MCF-7-CSC suspensions were adjusted to a density of 1 × 10 5 cells/ ml before cell SAP was separately added to a 6-well ultra-low adhesion culture plate at 2 mL/well, and cultured at 37°C with 5% CO 2 for 24 hours. Ad-VT was subsequently inoculated at 100 MOI. The proportion of CD44 + CD24 − cell subsets was detected by flow cytometry at 48 hours.

| Annexin V analysis
MCF-7-CSC suspensions were adjusted to a density of 1 × 10 5 cells/ mL before cell SAP was separately added to a 6-well ultra-low adhesion culture plate at 2 mL/well, and cultured at 37°C with 5% CO 2 for 24 hours. Then, Ad-VT, Ad-T, Ad-VP3 and Ad-Mock were subsequently inoculated at 100 MOI. At 24, 48 and 72 hours, the cells were collected by centrifugation at 2,000r/min for 5min before resuspension with 500 μL of 1 × Binding Buffer, followed by addition of 5 μL of FITC and 5 μL of PI. The cells were stained for 15-20 minutes in the dark, and then, 10 μL of cell suspension was absorbed and placed on a covered slide to be observed and photographed by confocal microscopy and fluorescence microscopy.

| JC-1 staining assay
JC-1 is used to detect the qualitative and quantitative changes of mitochondrial membrane potential (MMP). MCF-7-CSC suspensions were adjusted to a density of 1 × 10 5 cells/mL before cell SAP was separately added to a 6-well ultra-low adhesion culture plate at 2 mL/well and cultured at 37°C with 5% CO 2 for 24 hours. Ad-VT, Ad-T, Ad-VP3 and Ad-Mock were subsequently inoculated at 100 MOI. At 48 hours, the culture solution was discarded before the cells were treated with 1 mL of JC-1 dye diluted at 1:1000 (1 μL JC-1 and 1 mL culture solution), followed by 15 minutes of incubation in the dark, and then, 10 μL of sample was absorbed and placed on a covered slide to be observed and photographed by confocal microscopy and fluorescence microscopy.

| TMRM staining assay
MCF-7-CSC suspensions were adjusted to a density of 1 × 10 5 cells/ ml before cell SAP was separately added to a 6-well ultra-low adhesion culture plate at 2 mL/well and cultured at 37°C with 5% CO 2 for 24 hours. Ad-VT, Ad-T, Ad-VP3 and Ad-Mock were subsequently inoculated at 100 MOI. At 48 hours, the culture solution was discarded before the cells were treated with 1 mL TMRM dye at 10 μg/ mL, followed by 15 minutes of incubation in the dark. The cell slide was placed upside down to be observed by fluorescence microscopy.
After the above treatment, the cell samples were transferred into flow tubes for and analysis using flow cytometry.

| Statistical analyses
Statistical analyses were conducted using the data from at least three independent experiments using SPSS 20.0. P < 0.05 was used to indicate statistical significance. Data are presented as mean ± standard deviation (SD).

| Isolation, culture and identification of MCF-7-CSC
A cell cluster with proliferative ability was formed in the serumfree suspension culture of MCF-7. The volume of cell sphere was observed to have increased with prolonged culture time ( Figure 1A).
The proportion of CD44 + CD24 − cell subsets in the serum-free suspension culture of MCF-7-CSC was significantly higher than that of MCF-7 (P < 0.05) ( Figure 1B), indicating that the presence of surface markers of breast cancer stem cells in MCF-7-CSC had increased significantly than that of in MCF-7.
The expression levels of cancer stem cell marker proteins ALDH1A1, C-Myc, OCT4, NANOG, KLF 4 and SOX2 were found to be significantly higher in MCF-7-CSC than those in MCF-7 ( Figure 1C). Taken together, the results indicate that breast cancer stem cells MCF-7-CSC were successfully cultured.

| Characterization of MCF-7-CSC by tumourforming experiments in vivo
Mice inoculated with only 100 cells of MCF-7-CSC were seen to have developed tumours after only 30 days. Conversely, tumours were absent in mice inoculated with MCF-7 even at a 100-fold increased cell number of 1 × 10 4 . The tumour formation rate in mice with 500 MCF-7-CSC cells was 100%, whilst the tumour formation rate in mice with 1 × 10 5 MCF-7 cells was only 66.7% (Figure 2A-D). The results indicate that MCF-7-CSC in serum-free suspension culture had a stronger tumorigenic ability with certain stem cell characteristics.
After cell inoculation, whilst mice in the MCF-7-CSC group formed rapid-growing tumours, mice in the MCF-7 group formed fewer slow-growing tumours ( Figure 2E). The tumour volume doubling time of mice in the MCF-7-CSC group was significantly less than that of mice in the MCF-7 group (P < 0.05). The tumour multiplication time of mice inoculated with MCF-7-CSC showed a decreasing trend with the increase of inoculated cell number ( Figure 2E).
Consistently, the results show that MCF-7-CSC with certain stem cell characteristics were successfully cultured.

MCF-7-CSC inoculated with Ad-VT showed a significantly decreased
proportion of CD44 + CD24 − cell subpopulation in a dose-and timedependent manner, which was significantly different from that of the control group (P < 0.05) ( Figure 3B).

| Recombinant adenovirus induces cell apoptosis of MCF-7-CSC
In order to detect the cell apoptosis of MCF-7-CSC following infection with recombinant adenovirus, Annexin V staining was firstly performed. Confocal microscope images ( Figure 4A

| Recombinant adenovirus induces changes of mitochondrial membrane potential in MCF-7-CSC
To study the changes of mitochondrial membrane potential (MMP) in MCF-7-CSC, the cells infected with Ad-VT, Ad-T, Ad-VP3 and Ad-Mock, respectively, were stained with JC-1 staining solution at 24, 48 and 72 hours ( Figure 5A In agreement with our results above, Ad-VT exhibited the strongest effect in decreasing the MMP in MCF-7-CSC, followed by Ad-T and Ad-VP3, and there is no significant difference between Ad-Mock and control group (P > 0.05). In summary, the recombinant adenoviruses tested had a strong apoptosis-inducing effect on MCF-7-CSC cells, mainly through the mitochondrial apoptotic pathway.

| D ISCUSS I ON
The hypothesis of CSCs provides an important model for cancer research. 2 The key role of BCSCs in breast cancer initiation, metastasis and resistance highlights the urgent need for the development of BCSCs within 18 days, which was far more efficient than that with 1 × 10 5 MCF-7 cells. Taken together, our tumorigenesis experiment results showed that BCSC exhibited a very strong tumorigenic ability, which was significantly different from that of MCF-7.
As the source of tumour occurrence, development, recurrence and metastasis, CSCs are resistant to tumour treatment methods due to their high degree of self-renewal and differentiation ability. [28][29][30][31] BCSCs in breast cancer contribute to malignant progression; Annexin is a calcium ion-dependent phospholipid-binding protein that can specifically bind phospholipid serine (PS) inside the phospholipid bilayer of cells. When BCSCs undergo early apoptosis, the PS inside the everted cell membrane specifically binds Annexin V, which can then be detected by FITC green fluorescence. As the cell membrane remains intact, PI is inaccessible to the cell to stain the nucleus. However, at later stages of apoptosis, whilst the damaged cell membrane can be dyed green by FITC due to Annexin V binding, the nucleus can be stained red by PI. Although necrotic cells have no membrane phospholipid eversion, an increased cell membrane permeability enables PI to enter the cell to stain the nucleus.
Accordingly, BCSCs without undergoing apoptosis show no staining.
Under fluorescence microscope, BCSC apoptosis was observed after infection with Ad-VT, Ad-T and Ad-VP3. The cell membrane of apoptotic cells was stained green, whilst the nucleus was stained red with nuclear fragmentation. Consistently, flow cytometry analysis showed that Ad-VT exhibited the strongest ability to induce apoptosis, followed by Ad-T and Ad-VP3. The apoptosis rates were observed to be increased in a time-dependent manner. Taken together, our results indicate that Ad-VT mainly kills BCSC through apoptosis.
Next, we analysed changes in the mitochondrial membrane potential (MMP) using fluorescent probe JC-1. In normal cells, the high MMP enables the JC-1 to enter the mitochondrial matrix where a polymer is formed to emit red signal at a high concentration.
Conversely, when the cell undergoes apoptosis, decreased MMP prohibits the JC-1 from entering the mitochondria to form a polymer, so JCI remains in the cytoplasm as a monomer that emits green signal. Accordingly, the apoptosis can be detected by observing the fluorescent colour of JC-1.
After 72 hours of infection with recombinant adenovirus, normal cells showed red signals, whilst apoptotic cells showed green signals following JC-1 staining. Ad-VT, Ad-T and Ad-VP3 were able to induce apoptosis of BCSCs by decreasing the MMP of apoptotic cells to be stained green by JC-1. Compared to cells infected by Ad-T and Ad-VP3, a higher rate of apoptosis was observed in cells infected F I G U R E 5 Effect of Ad-VT on the mitochondrial membrane potential of MCF-7-CSC cells. (A-B) Changes in red and green fluorescence signals measured by fluorescence microscopy after JC-1 staining. Increased apoptosis resulted in a decrease in the ratio of red to green fluorescence. Quantitative measurement of changes in the ratio of red to green fluorescence after JC-1 staining. Ad-VT clearly altered the mitochondrial membrane potential (MMP). Ad-VT had the strongest ability to induce apoptosis by affecting the MMP. (C-D) The MMP of MCF-7-CSC cells was analysed by fluorescence microscopy and flow cytometry after TMRM staining. The MMP of MCF-7-CSC cells was significantly decreased after infection with Ad-VT. The scale bar equals 50 μm. Data are shown as mean ± SD (*P < 0.05, **P < 0.01 and ***P < 0.001) when compared with control by Ad-VT. In agreement with the results obtained by JC-1 staining, quantitative analysis of MMP changes by TMRM staining showed that MMP decreased most significantly in Ad-VT infected cells.
In summary, we successfully isolated and cultured breast cancer stem cells, and found that recombinant adenovirus Ad-VT had a killing effect on BCSCs, and that the killing effect on BCSCs was mainly caused by apoptosis. Our results provide a new theoretical basis for the future treatment of breast cancer.

ACK N OWLED G EM ENTS
This work was supported by the National Science and Technology major Project (grant number 2018ZX10101003-005-003).

CO N FLI C T O F I NTE R E S T S
The research was conducted in the absence of any commercial or financial relationships that could be deemed as a potential conflict of interest.