Quantified The Tumor Cell Splits And Isolated The Cell Subpopulations Alive For Analysis In Vitro And In Vivo With The Vibrant Fluorescent Dye

Background and ObjectiveFor decades, tumor heterogeneity has remained a significant challenge for cancer therapy. Individual cancer cells behave differently in terms of cell proliferation, metastasis and sensitivity to therapy. Increasing evidences suggest that only a small proportion of cells, considered as cancer stem cells (CSCs), are responsible for tumor initiation, metastasis and resistance to chemotherapy due to their properties of self-renewal and differentiation into multiple cell types. Therefore, the identification and isolation of CSCs is of the upmost importance in order to cure cancer.To study the behavior of cancer stem cells, we are supposed to optimize the method of evaluating cell proliferation. Many alternative techniques are available to investigate cell proliferation both in vivo and in vitro. The cell proliferation is only amenable to be studied with these techniques which examine cell division in vitro culture over a narrow time window, or can trace cells which have divided recently, but not allow the further analysis of recovered viable cells, nor quantitatively measure the divisions of specific sub-groups.Traditionally, proliferative assays have been performed through incorporation of 5-Bromo-2’deoxyuridine (BrdU), which substitutes for thymidine in the DNA of dividing cells. The disadvantages are that comparatively high concentration of BrdU must be present throughout the period of cell split and only three consecutive cycles can be tracked. Besides, in some cell types, cells may accumulate in G2 phases after labeled with BrdU, and viable cells are not recoverable for further functional studies after NP4 treatment. Samples need to be denatured in Brdu immunochemical staining method, which will influence the structure of the samples and make it unsuitable for the downstream experiments. Ross et al reported that a low-dose, single-pulse of BrdU exhibit a dramatic effect of anti-proliferation in cultured neural progenitor and stem cells. This was coupled with changes of cell characteristics and senescence induction by BrdU.Another assay applied for determining cell proliferation in vitro is the MTT test. The yellow tetrazolium salt MTT could be metabolized into blue formazan crystals by activation of mitochondrial dehydrogenases in viable cells, exclusively. Then the crystals are dissolved by DMSO and detected by spectrophotometrical analysis.MTT assay has been commonly used to measure the potential cytostatic activity or cytotoxicity of medicinal agents, whereas it is not sensitive enough for a precise quantification of proliferation.The fluorescent dye carboxyfluorescein succinimide ester (CFSE) has been commonly used for tracking cell division, migration and positioning both in vivo and in vitro, as the fluorescence intensity decreases to half with every cell division. Approximately 8 cell divisions can be tracked before the CFSE fluorescence declines to the autofluorescence and cannot be distinguished. However, one of the limitations is that CFSE can be toxic for cells and inhibits cell proliferation to some extent. Although reducing the concentration of CFSE avoids this problem but also limits the tracking for more cell divisions.There were studies applying of PKH label retention/quenching to identify quiescent cells in glioblastoma, breast cancers and ovarian adenoma. They used terms, such as high, low and non-expression, to describe the properties of the sorted cells, but these terms were subjective. It is a complicated and time-consuming method requiring the addition of various reagents and several washing steps labeling with PKH26. Though PKH26 can be used to trace cells both in vivo and in vitro, there is a limitation for evaluation by frozen tissue sections and special mounting techniques.In this artical, we described the technique of staining cells with DiO/DiD dyes for tracing tumor cell proliferation and quantifying the cell splits numbers in vitro and in vivo. The long-chain carbocyanine dyes are essentially insoluble and weakly fluorescent in aqueous solutions. When cells are cultured in dilute dye solutions, dyes are incorporated into cells with their alkyl chains embedded in the lipid bilayer and become highly fluorescent and reasonably photostable. This study describes the applications of DiO/DiD for tracing cell splits based on the successive halving with outstanding fidelity of this fluorescent vital dye. When the labeled cells divide, the membrane fluorencence dyes are separated into daughter cells. Except for the low fluorescence variance, high stability of cell labeling, low cytotoxicity, and high resistance to intercellular transfer, the capability of recovering viable cell progeny and monitor the kinetics of proliferation are major advantages, as functional studies can then be undertaken. This technique could potentially contribute to gather information on cell proliferation kinetics and cell differentiation, accompanied by concurrent analysis of other characteristics, such as the expression of surface molecules and protein.Using the DiO/DiD labeling technique, we also explore the effect of hepatic lipase (LIPC) and CD 133 on the cell proliferation and chemotherapy in HepG2 cells. In the preliminary studies of our research group, we have found that down regulation of LIPC in HepG2 cells was associated with decreased cell proliferation and colony formation, decreased expression of CD 133, as well as increased resistance to chemotherapy. In this study, the dye labeling technique was used for further investigation of the correlation between LIPC, CD 133 and tumor response to chemotherapy.Methods1. Investigate the feasibility of DiO/DiD labeling for monitoring cell proliferation, including the cytotoxicity and stability of DiO/DiD dye.1.1 MTT Assay for DiO labeled cells proliferation. After being labeled with DiO dye, the cell proliferation of Lovo, SW480, SW620, CT26 and HepG2 cells were determined by MTT assay, respectively.1.2 Testing of the stability of DiO/DiD labeling both in vitro and in vivo.1.2.1 Co-culture assay in vitro. A co-culture assay with DiO-labeled human cell line HepG2 (EPCAM positive) and DiO-unlabeled murine cell line CT26 (EPCAM negative) was performed at a mixture of 1:1 for 7 days in vitro. Cells were collected and followed with anti-human EPCAM antibody incubation and FACS detection.1.2.2 DiD labeling in vivo. Female 5-to 6-week-old BALB/c nude mice were implanted in liver with DiD positive HepG2 cells. On the 8th day, tumors were collected and single cell suspensions were obtained. Thereafter, cells were incubated with anti-human EPCAM antibody and the fluorescence intensities of DiD dye were analyzed.2. Establish the standard curves of cell splits with DiO labeling both in vitro and in vivo.2.1 Principle:An equation was proposed based on the fact that each time a cell divide, the fluorescence intensity of DiO dye was reduced by half. Presumptively, Ma was the initial average fluorescence intensity of each cell; Na was the initial cell number; Mb was the terminal average fluorescence intensity of each cell; Nb was the terminal cell number; S was split number; then we may conclude that:Ma* Na=Mb*Nb, log2 (Ma/ Mb)=S=log2 (Nb/Na), S= log2 Ma-log2 Mb .By plotting the fluorescence intensity values versus cell split numbers, we carried on the logarithmic fitting curve according to R programming language and Excel software.2.2 The standard curve of cell proliferation cultured in different concentration of serum. After stained with DiO, cells were cultured in 6-well microplates in different concentration of serum(10%、5%、2.5%、1.25%、0.625%). The cells were harvested when the cells reached 80%-90% confluence in different time points and then cell numbers were counted using a hemo-cytometer and the intensities of DiO fluorescence were analyzed by FACS. A standard curve was generated by plotting fluorescence values versus cell split numbers.2.3 The standard curve of cell proliferation in different time point. After labeled with DiO, HepG2 and CT26 cells were cultured in 6-well microplates and were harvested at a confluence of 80-90%.4/5 of the cells were counted and analyzed by FACS. The remaining 1/5 cells were passaged. The above procedure was repeated until the DiO fluorescence intensity decreased to the autofluorescence level of the unstained active cells. The total numbers of cells were calculated. A standard curve was generated by plotting fluorescence values measured in these samples versus cell split numbers.3. Applications of DiO labeling technology both in vitro and in vivo.3.1 The effect of chemotherapy on DiO-labeled HepG2 cells. DiO labeled HepG2 cells were cultured for ten days under low concentration of doxorubicin (0.1 μM, 0.3uM) and then detected by flow cytometry. Cells had separated into two distinct groups, which were identified and isolated as the DiO-Low population and DiO-High population by virtue of their ability to retain the DiO labeling. The percentages of cancer stem cell biomarker CD 133+ were determined in HepG2 cells treated with OuM and 0.3uM doxorubicin after they were incubated with anti-human APC-labeled CD 133 human antibody.3.2 DiO labeling contributed to tumor cell subgroup sorting. The DiO-Low population and DiO-High population were isolated by FACS and sequentially cultured for three days. Then ribosome nascent-chain complex-bound mRNAs (RNC-mRNAs) were extracted from both populations. RNC-mRNAs were considered to be highly correlated the relative abundances of proteins rather than total RNA. To extract the RNC-mRNAs, cells were pre-treated with 100 mg/ml cycloheximide at 37"C for 15 min, then washed by pre-chilled PBS for twice and added 1 ml cell lysis buffer. After 30-min ice-bath, scrape the cell lysates and transfer to pre-chilled 1.5 ml EP tubes. Centrifuge at 16000 g for 10 min at 4℃ and remove the cell debris. Transfer the supernatants on the surface of 8 ml of 30% sucrose buffer. RNCs were pelleted after ultra-centrifugation at 330000 g for 3 h at 4 ℃. The RNCs pellets were dissolute with 100 ul RB buffer and 1 mL Trizzol was added. Then RNA was extracted in routine method.3.3 cDNA microarray analysis. Microarray experiments were performed by RiboBio Co., Ltd (Guangzhou, China). Initially, RNA of the DiO-Low population and DiO-High population was extracted and the quantity of RNA was determined with Agilent 2200 Bioanalyzer (Agilent, USA). Three independent samples of each group were prepared and equal amount of total RNA from each preparation was pooled, respectively. The CDNA was synthesized and labeled with Cy3/Cy5, and randomly hybridized to a RiboArrayTM Custom Array 1x40K (RiboBio) according to the manufacturer’s instructions. The hybridized microarray was scanned and the data were analyzed and normalized. The differentially expressed genes were identified at a fold change≥2.3.4 Cell apoptosis assay of DiO labeled cells. HepG2 cells were treated with 0.3 uM doxorubicin and cultured for a week, then treated with doxorubicin at a concentration of 20uM in complete medium and incubated at 37℃ in a 5% CO2 incubator for 4 hours to induce apoptosis. Remove the culture medium and wash twice with PBS. Add fresh complete medium and incubate for another 12 hours. Cells were collected and apoptosis assay was performed according to the kit instructions. The fluorescence intensities of DiO in viable cells population and the early apoptotic cells population were determined, respectively, and cell split numbers were calculated for analysis.3.5 Application of DiO labeling in vivo experiments.3.5.1 Subcutanious implantation of DiO labeled CT26 cells. DiO-labeled CT26 cells suspension (about 2*106 cells/mouse) were implanted subcutaneously into the dorsal region of 5-8-week-old female BALB/c nude mice. Then remove the primary tumors and collect the cells to obtain single cell suspension on the day 3,9,14, respectively. The fluorescence intensities of DiO were detected by FACS. According to the fitting formula, the cell split numbers were calculated.3.5.2 Spleen capsule implantation of DiO labeled CT26 cells. DiO-labeled CT26 cells were injected into mice subcutaneously or the spleen capsule. Three days later, tumors in situ and the liver metastases were collected and the fluorescence intensities of DiO were detected. According to the fitting formula, the cell split numbers were calculated.3.5.3 Liver capsule implantation of DiO-labeled CT26 cells. Mice were injected with DiO-stained CT26 cells into the liver capsule and were sacrificed on the day 3,4, 5,9,10 after transplantation, respectively. The tumors in the injection area and the metastases in other liver lobes were collected. The fluorescence intensities of DiO were detected and the cell split numbers were calculated.4. DiD labeling for exploring the effect of LIPC on chemotherapy in HepG2 cells.4.1 HepG2 cells were transfected with LV-LIPC-shRNA-EGFP, without selecting the positive clones. (It meant that the transfected cells contained both EGFP+cells and EGFP-cells.)4.2 The transfected HepG2 cells were labeled with DiD dye and treated with 0.03, and 0.3μM doxorubicin for ten days. Then the DiD intensities were detected by FACS.5. Statistical analysesStatistical analyses were performed using the SPSS 20 software. All data were presented as mean±SD of three independent experiments. For cell functional assays, the statistical significance between DiO-Low and DiO-High groups were determined by two-tailed unpaired Student’s Mest. For chemoresistance experiments, data were analyzed with two-factor variance analysis and SNK (Student-Newman-Keuls) test. P values<0.05 were considered statistically significant.Results1. The effect of fluorescent dyes DiO labeling on cell proliferarion.MTT assay revealed that there was no significant difference between labeled cells and unlabeled cells in the day 1,2,3 (P>0.05), demonstrating that cell proliferation was not affected by DiO labeling.2. The station of DiO/DiD labeled cells in vitro culture and in vivo culture.A co-culture assay with DiO labeled human cell HepG2 (EPCAM positive) and DiO unlabeled murine cell CT26 (EPCAM negative) was performed and the results showed that few DiO positive cells were detected in EPCAM negative CT26 cells(8%±0.096%), indicating that the dyes rarely transferred from labeled cells to unlabeled cells in vitro culture. Eight days after subcutaneous transplantation of DiD negative or DiD positive HepG2 cells into mice liver capsule, the primary tumors were collected and single-cell suspensions were obtained followed with anti-human EPCAM antibody incubation and FASC detection. Data showed that few DiD positive cells were detected in EPCAM negative cells (4.5%±0.08%), indicating that the dyes rarely transferred from labeled cells to unlabeled cells in vivo culture.3. Establish the standard curves by plotting fluorescence intensity values versus the cell split numbers.3.1 The standard curve of cell proliferation cultured in different concentration of serum was generated by plotting fluorescence values versus cell split numbers. The fitting formulas were:Lovo:S=-1.79308log2M+16.41884, R2=0.7871; SW480: S=-2.6156 log2M+16.37847, R2=0.9033; HCT116:S=-2.77449log2M+29.27126, R2=0.8867; HepG2:S=-1.4299log2M+18.82296, R2=0.9044; SW620: S=-1.71533log2M+13.34059, R2=0.9853.3.2 The standard curve of cell proliferation in different time point was generated by plotting fluorescence values versus cell split numbers. The fitting formulas were: CT26:S=-1.28089log2M+14.00148, R2=0.9941; HepG2:S=-1.0812 log2M+11.56886, R2=0.9963. Data indicate that the fluorescence intensity of DiO decayed in the time window of cell splits had no effect on the correlation of the splits and the fluorescence.4. The effect of chemotherapy on DiO-labeled HepG2 cells.DiO-labeled HepG2 cells had been cultured for ten days and were identified and isolated into two sub-group of cells by FACS as the DiO-High (split-slow population) and the DiO-Low (split-fast population). The percentage of CD 133+cells in DiO-High population were increased significantly compared with DiO-Low population (P<0.05). We concluded that DiO-High population was more resistant to chemotherapy.5. Comparison of DiO intensities of early apoptotic cells and living cells.A standard curve of cell division of HepG2 cells treated with 0.3 uM doxorubicin was established and the fitting formula was determined as:S=-1.59485log2M+17.407, R2=0.9202. According to the fitting formula, we calculated that the early apoptotic cells population split 3.13-times averagely and the all living cells population split 2.96-times averagely (P=0.035). The result indicated that cells divided at a faster rate were more sensitive to doxorubicin and underwent the apoptosis.6. Cell sorting and quantitative analysis of DiO-High/Low HepG2 cells.DiO-High/Low subpopulations were isolated by FACS and the fluorescence intensities were measured. According to the fitting formula S=-1.594851og2M+17.407, R2=0.9202, we calculated that DiO-Low population split 9.6-times averagely and significantly different from the DiO-High population, split 2.9-times averagely, respectively.7. cDNA microarrays of DiO-High/Low HepG2 cells.. The gene expression profiles were screened using ribosome combining cDNA with cDNA microarrays. A total of 6521 differentially expressed genes were identified in these two sub-populations, in which 2768 of them were up-regulated and 3753 were down-regulated. Gene ontology (GO) term enrichment analysis revealed that the genes involved in regulation of cellular process, cellular response to stimulus and biological regulation were among the most severely affected in response to different speeds of cell proliferation. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that the enrichment of biological signal pathway include metabolic pathways, MAPK signaling pathway, cAMP signaling pathway, PI3K-Akt signaling pathway, AMPK signaling pathway and Pathways in cancer. The gene expression in cell apoptosis, cell cycle and pluripotent stem cell pathway were significantly different between DiO-High and the DiO-Low (split-fast population). Data showed that some vital pro-apoptotic genes, including BIM, BID, BAD, FAS, PRKACB, CASP7, CASP8, CASP10, CASP12 were downregulated, while some significant pro-survival genes, including DFFA, RELA, NTRK1, CFLAR were upregulated in DiO-High population. The result indicated that cells divided at a faster rate were more vulnerable to doxorubicin, which was capable of inducing apoptosis. We screened out several genes contributed to cell growth (PCNA、CDK4s cyclin D2、cyclin El、cyclin Al、cyclin B3) and found that they were down-regulated in DiO-High population compared with DiO-Low population. Moreover, in pluripotent stem cell pathway, some stem cell related genes were significantly upregulated in DiO-High population, indicating that DiO labeling could be a useful tool for tumor subgroup sorting and help intense researches of tumor heterogeneity and potential therapeutic targets.8. In vivo culture and quantitative analysis of DiO labeled tumor cells.8.15-8-week-old male mice were implanted subcutaneously with DiO-labeled CT26 cells. According to the fitting formula S=-1.280891og2M+14.00148, we calculated that the cell split numbers in the day 3,9,14 were 1.2±0.072,2.59±0.078,5.80±0.17, respectively. In vivo, the level intensity of DiO fluorescence was significantly decreased while the cells split over times.8.2 DiO-labeled CT26 cells were injected into mice subcutaneously or the spleen capsule. Three days later, tumors in situ and the liver metastases were collected. According to the fitting formula S=-1.280891og2M+14.00148, we calculated that CT26 cells split 1.2±0.072 times in subcutaneous,5.88±0.20 times in spleen and 7.78±0.30 times in liver metastases. CT 26 cells split 3.6±0.41 times in spleen more than under skin, averagely. Moreover, in liver metastases, CT26 cells split 1.2±0.15 times faster than the cells in primary tumors in spleen, averagely. We found that tumor cells grew faster in spleen than in subcutaneous tissues and the intensity of DiO in liver was the lowest, indicating that cell division in metastatic tumor was faster than tumor in situ.8.3 Mice were injected with DiO-stained CT26 cells into the liver capsule and were sacrificed on the day 3,4,5,9,10 after transplantation, respectively. The tumors in the injection area and the metastases in other liver lobes were collected. According to the fitting formula S=-1.280891og2M+14.00148, we calculated that the cells split numbers were 2.14±0.30,3.61±0.73 on the third day; 2.67±0.44,3.99±0.22 on the fourth day; 5.69±0.44,7.86±0.33 on the fifth day; 6.92±0.26,7.63±0.07 on the ninth day; 6.86±0.96,9.15±0.40 on the tenth day. Data showed that there was a significant difference of DiO fluorescence intensity between the in situ and metastasis tumors. In the same organ, the metastatic CT26 cells split about 1.1 times faster than the primary CT26 cells in the same intervals, averagely.9. Explore the effects of hepatic lipase on cell proliferation and chemotherapy of HepG2 cells with DiD labelding technique.The HepG2 cells transfected with LV-LIPC-shRNA-EGFP (without positive selection) were labeled with DiD and treated with 0.03,0.3μM doxorubicin. Ten days later, the fluorescence intensity of DiD was determined. The results showed that the EGFP(+) cells divided more slowly than the EGFP(-) cells, while the EGFP(-) cells had separated into two distinct subpopulation according to the DiD intensities. Considering the previous research results, that knocking down LIPC in HepG2 cells inhibited cell proliferation and colony formation, together with the decreased CD133 expression and increased resistance to chemotherapy.ConclusionFrom the results above, we may come to the following conclusions:1. Fluorescent dyes DiO/DiD did not affect tumor cell proliferation and rarely leaked out from cells.2. A standard curve of cell division could be established by plotting fluorescence intensity values versus cell split numbers, and the fitting formulas could be used to quantified tumor cell splits both in vitro culture and in vivo culture. And the fluorescence intensity of DiO decayed in the time window of cell splits had no effect on the quantitative analysis.The experiment in vivo showed that cell division in metastatic tumor was faster than tumor in situ.3. DiO labeled cancer cells separated into two distinct sub-groups after chemotherapy, and the DiO-High population showed increased expression of CD 133 and more resistant to chemotherapy.4. The apoptosis assay verified that DiO-Low (split-fast) population was more likely to go on apoptosis and more sensitive to chemotherapy.5. cDNA microarray verified that in DiO-High (split-slow) population, some vital pro-apoptotic genes and pro-proliferation genes were downregulated, while some significant pro-survival genes and stem cell related genes were upregulated.6. Down regulation of LIPC in HepG2 cells was associated with decreased cell proliferation and colony formation, decreased expression of CD 133, as well as increased resistance to chemotherapy.