RSH-transformed cells acquire cancer cells characteristics

Primary human fibroblast cells IMR90 were successfully co-transfected with Ras, SV40 Large-T, and hTERT and their protein expressions were confirmed by western blotting (Figure 1A). Morphologically, IMR90 RSH fibroblasts appeared to be shorter and rounder compared to the infection control (Figure 1B). This observation is consistent with the findings of Mason and colleagues in IMR90 cells transformed with E1a/Ras [17], suggesting that these changes are the unique characteristics of cellular transformation. Moreover, late passages of IMR90 control cells underwent significant increase in cell sizes, indicating their senescent status. However, this was not observed in IMR90 RSH cells even after several passages (data not shown).

One of the hallmarks of the cancerous cells is they can survive and grow in the absence of anchorage to the extracellular matrix [18]. Our anchorage independent growth assays demonstrated that IMR90 RSH cells formed small microscopic colonies (<200 μm in diameter) while MCF-7 cells, the positive control, formed large visible colonies (>200 μm in diameter) (Figure 1C) after 6 weeks. Comparison of colony sizes with MCF-7 suggests that transformation by three genetic factors produced cells that were less tumorigenic than the established cancer cells. Thus, the RSH-transformed cell could serve as a representative model to study the early events of cancer transformation, compared to an established cell line.

Another common trait of cancer cells is their self-sufficiency in growth signals [18]. To investigate the effects of growth factors withdrawal on the RSH-transformed cells, the cells were subjected to a serum-free environment, following which cell proliferation and survival were assessed over a three-day period. As expected, IMR90 control cells showed signs of apoptosis after 24 hours of serum withdrawal. However, no apoptosis was observed in IMR90 RSH cells even after treating for 72 hours in serum-free condition, suggesting that these RSH-transformed cells were able to survive in the absence of growth factors (Figure 1D). We further tested whether transformed fibroblasts are refractory to the induction of apoptosis, which is commonly observed in cancer cells [19]. After treating with Doxorubicin (Dox), a DNA damage-inducing drug for 48 hours, IMR90 RSH cells were able to survive even at much higher Dox concentrations (3 μM and 5 μM) than IMR90 control cells, reflecting an attenuation in the apoptotic machinery of the cells (Figure 1E).

For potential use of IMR90 RSH as a cancer cell model, we also assessed their chromosomal aberrations for recurrent abnormalities through cytogenetics study [20]. Karyotyping of IMR90 RSH cells revealed recurrent abnormalities (Figure 1F,G). Moreover, spectral karyotyping analysis revealed that 60.9% (14 out of 23) of IMR90 RSH cells had recurrent chromosomal abnormalities at chromosomes 4, 18, 20 (Figure 1H; Additional file 1: Table S1). Taken together, these data indicated that IMR90 cells were successfully transformed by co-expressing three oncogenic factors. The RSH-transformed fibroblasts acquired various characteristics of a human cancer cell and may serve as a valuable model to study the early events of tumorigenesis.

RSH-transformed cells and hTERT-overexpressing cells demonstrate increased migration capability

Metastasis is often correlated to two attributes: migration followed by invasion [21]. Metastatic tumors are believed to occur in the late stages of cancer [22]. However, in numerous early-stage cancers, recurrence can be observed even after the removal of non-invasive, benign tumors, suggesting the possibility of the cancer cells undergoing metastasis at a much earlier stage [23]. We questioned if IMR90 RSH cells possess migration capability similar to that of a cancer cell. Our Boyden assay results showed that IMR90 RSH cells had a greater migration capability as compared to IMR90 control cells (Figure 2A). This result was further validated with wound healing assay, in which the gap was reduced to 40% for IMR90 RSH cells compared to 84% in IMR90 control cells (p < 0.01) (Figure 2B,C). Moreover, we observed that IMR90 RSH cells migrated faster and in a more individualistic pattern compared to IMR90 control cells, implying some degree of autonomy (Figure 2B).

Given that transformation can increase the migration capability of cells and that hTERT is one of the upregulated factors in the transformed cells, it then raised the question as to whether hTERT alone can contribute to this phenomenon. In order to assess the possible role of telomerase in cell migration, we also performed wound healing assay on IMR90 cells expressing hTERT alone. Similar to IMR90 RSH cells, the hTERT-overexpressing IMR90 cells (Figure 2D) also migrated faster than IMR90 control cells (Figure 2E). However, when compared to IMR90 RSH cells, the migration process in IMR90 hTERT cells started to occur only after 10 hours and they seemed to have a slower migration rate. As shown in Figure 2E and F, IMR90 hTERT cells migrated about 50% distance, yet the distance was only reduced by 10% for IMR90 control cells (p < 0.01). Collectively, these suggest that hTERT does play a role in cell migration, which in turn contributes to the metastatic potential of cancer cells.

Microarray analysis supports the notion that hTERT plays a role in cell migration

On the other hand, Liu et al. in 2013 have shown that hTERT promotes the epithelial-mesenchymal transition (EMT) by upregulating snail family zinc finger 1 (Snail1) and vimentin through Wnt/beta-catenin signaling pathway in gastric cancer [26]. We did not observe significant changes in Snail-1 and vimentin genes expressions in our cell models and this could be attributed to different cell model used in both studies. However, we noticed that two downstream genes of Wnt/beta-catenin pathways, namely Cyclooxygenase-2 (COX-2) and Twist1 were significantly enhanced in both RSH-transformed cells and hTERT-overexpressing cells. Both were reported to be implicated in cancer cell migration [27, 28], and this is in consistent with our microarray data. Taken together, our microarray data complements with our previous notion that hTERT play an important role in cell mobility, which in turn contributes to the metastatic potential of cancer cells.

hTERT might be implicated in DNA damage response via upregulating Ku70 during IMR90 transformation

To delineate the roles of hTERT in neoplastic transformation, we performed mass spectrometry analysis on the protein levels of IMR90 RSH cells. Surprisingly, analysis by mass spectrometry revealed that protein Ku70 was exclusively found in IMR90 RSH cells, but not in IMR90 control cells (Figure 3A; Additional file 2: Table S2). Moreover, both RT-PCR and immunoblotting results confirmed our observation where Ku70 expression was augmented in IMR90 RSH as well as IMR90 hTERT cells, but not in IMR90 control cells (Figure 3B,C). This could be indicative of an activated DNA damage responses (DDR) in both IMR90-RSH and IMR90-hTERT since Ku70 is a DDR sensor and is implicated in the non homologous end joining (NHEJ) pathway [29]. Ku70 forms a heterodimeric complex with Ku80 [30]. However, the protein expression of Ku80 remained unchanged (Figure 3C). Presence of Ku70 in IMR90 RSH cells and IMR90 hTERT cells prompted us to assess whether other DDR-associated proteins could be regulated by hTERT as well. Our microarray data revealed that several DDR-associated genes were upregulated in both RSH-transformed IMR90 and hTERT-overexpressing IMR90 cells (fold change ≥ 2; p < 0.05) (Additional file 3: Table S3). Taken together, these results suggest that hTERT may play an important role in DDR pathways, via Ku70 and other DDR-associated proteins.

Cell lines

Human breast cancer cells MCF-7 (HTB-22) and normal human fibroblasts, including IMR90 (CCL-186) and BJ (CCL-186) were purchased from American Type Culture Collection (ATCC). Cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Gibco, Invitrogen), 2 mM L-glutamine (Gibco, Invitrogen) and 100U/ml penicillin and streptomycin (Gibco, Invitrogen).

Infection and selection of cells

Retroviral pBabe-puro Ras V12 (Addgene Plasmid 1768), and retroviral pBabe-puro SV40 LT (Addgene Plasmid 13970) were purchased from Addgene. FLAG was tagged to the N-terminus of hTERT and cloned into a lentiviral vector with Ires hygromycin mammalian selection. Retroviral pBabe-puro-hTERT vector (Addgene Plasmid 1771) was used to generate hTERT-overexpressing cell. Retroviral pBabe-puro vector (Addgene Plasmid 1764) was used as infection control. IMR90 was infected with virus supernatant with 8 μg/ml of polybrene for 24 hours. To generate stable cell line that co-expresses Ras, SV40 Large T and hTERT, cells infected with RSH combination were selected using both puromycin and hygromycin (Invitrogen) for 5 days. Cells infected with pBabe-puro-hTERT as well as pBabe-puro control vector were selected using puromycin (InvivoGen) for 5 days. Cells stably expressing the desired genes were further passaged and maintained on selection medium for an additional two to three weeks prior to downstream experiments.

Anchorage-independent growth assay

10⁴ cells were seeded in 0.3% (w/v) agarose with DMEM and 16.66% FBS onto each well of a 24-well plate with a 0.6% (w/v) agarose underlay. Wells were analyzed for colony formation after 6 weeks. Scoring was done by counting the colonies under the microscope.

Serum-free cell survival assay

10⁵ cells were seeded and the number of living cells was determined by Countess automated cell counter (Invitrogen) after treating them in serum-free DMEM for 24, 48 and 72 hours.

Drug treatment

10⁵ cells were treated with Doxorubicin (Calbiochem) at 1 μM, 3 μM and 5 μM concentrations. The number of living cells was determined by cell counting 24 and 48 hours after drug treatment.

Immunoblotting

Immunoblotting was performed as described previously [31]. The following antibodies were used: H-Ras (Santa Cruz), SV40 Large-T (Santa Cruz), FLAG-Tag (Sigma), hTERT (Epitomics), Ku 70 (Cell Signaling), Ku 80 (Cell Signaling), α-tubulin (Sigma) and β-actin (Abcam).

Wound healing assay

6-well plates were first coated with 10 μg/ml of collagen (Cohesion) for two hours, followed by blocking with BSA for one hour. 5 × 10⁵ cells were seeded into the wells and allowed to grow until confluent. Following which, the cell monolayer was scratched at the bottom of the wells. Culture medium was then replaced with serum-free DMEM to minimize cell proliferation. Wells were observed under the microscope at different time-points and the distance of gap was measured. The percentage of wound closure was calculated from distances quantified in three independent experiments.

Boyden assay

Migration assay was performed using cell culture inserts (BD Biosciences) with 8 μm pore size. 10⁵ cells were diluted in 300 μl of serum-free DMEM supplemented with 0.1% (w/v) BSA and added to the upper chamber of the well. 500 μl of normal culture DMEM supplemented with 100 ng/ml of human epidermal growth factor (ProSpec) was added to the lower chamber and incubated for 10 hours. Following which, medium was removed from both upper and lower chambers and cells were removed by swabbing from the upper chamber. Cells from the lower side of the insert were then stained with 1% crystal violet and observed under the microscope.

Illumina microarray

Total RNA was extracted using RNeasy Mini Kit (Qiagen). 500 ng total RNA was used for cRNA amplification using Illumina TotalPrep RNA Amplification kit (Ambion), following manufacturers’ instructions. 750 ng cRNA was then hybridized onto the HumanRef-8 v2 Sentrix BeadChip (Illumina). Subsequently, the fluorescence emission by Cy3 was quantitatively detected by the Illumina BeadArray Reader software for downstream analysis of data by Partek software. Statistical significance of individual gene expression levels was analyzed by Analysis of variance (ANOVA) and significant genes were identified based on fold-changes. Threshold for significance was set at fold-change > ± |2.0| when compared to control cells. Absolute intensity differences between experimental groups were set at p < 0.05. All data is MIAME compliant and the raw data has been deposited in a MIAME compliant database (accession number: GSE24097).

RT-PCR and DNA gel electrophoresis

cDNA was used in a total of 25 μL PCR reactions containing PCR buffer, DNA Taq polymerase, 200 nm of each primer and dNTP. PCR products were resolved using gel electrophoresis.

Metaphase spreads

Metaphase spreads were prepared as described in Jeppesen’s protocol [32], with slight modifications. Slides were observed using the Olympus Fluoview 1000 confocal microscopy system.

G-band karyotyping and spectral karyotyping

Cells at about 80% confluence were treated with colcemide for mitotic arrest and harvested by standard hypotonic treatment and methanol: acetic acid (3:1) fixation. For G-band karyotyping, slides were prepared by standard air drying method and G-band karyotype was performed according to the published protocols. For spectral karyotyping, slides were prepared by standard air drying method and hybridized with Human SKY paint probe (ASI), as per manufacturers’ recommendations. 23 metaphases were analyzed by SKY analysis. Identified chromosomal abnormalities were described according to the International System for Human Cytogenetic Nomenclature (ISCN) (1995). Recurrent abnormalities are defined as at least 3 metaphase cells having the abnormality at the same region of chromosomal location or those that are involved in > 3% of cases.

Coomassie blue staining and mass spectrometry

Coomassie blue staining was performed on 7% and 15% SDS-PAGE gel. Bands were excised at 70 kDa and the spots were rehydrated in digestion buffer containing sequencing grade modified trypsin at 37°C. Digested peptides were extracted from gel with TFA extraction buffer and desalted using C-18 Zip-tips (Millipore). Mass spectra of the peptides in each sample were obtained by MALDI-TOF. Protein identification was based on peptide fingerprint mass mapping and peptide fragmentation mapping.

Availability of supporting data

The data sets supporting the results of this article are available in the LabArchves. The unique persistent identifier and hyperlink to data sets are listed below:

http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE24097 (Microarray data)

https://mynotebook.labarchives.com/share/Xu%2520Cao/MjAuOHw0MDkwNC8xNi9UcmVlTm9kZS84MDAxMjY3Mjd8NTIuOA== (Additional file 1: Table S1)

https://mynotebook.labarchives.com/share/Xu%2520Cao/MjIuMXw0MDkwNC8xNy9UcmVlTm9kZS8zMjY0NzI2ODU1fDU2LjE= (Additional file 2: Table S2)

https://mynotebook.labarchives.com/share/Xu%2520Cao/MjMuNHw0MDkwNC8xOC9UcmVlTm9kZS80MTE5NjA3NzE4fDU5LjQ= (Additional file 3: Table S3)