ABSTRACT
Previous observations that human amniotic fluid cells (AFC) can be transformed by human adenovirus type 5 (HAdV-5) E1A/E1B oncogenes prompted us to identify the target cells in the AFC population that are susceptible to transformation. Our results demonstrate that one cell type corresponding to mesenchymal stem/stroma cells (hMSCs) can be reproducibly transformed by HAdV-5 E1A/E1B oncogenes as efficiently as primary rodent cultures. HAdV-5 E1-transformed hMSCs exhibit all properties commonly associated with a high grade of oncogenic transformation, including enhanced cell proliferation, anchorage-independent growth, increased growth rate, and high telomerase activity as well as numerical and structural chromosomal aberrations. These data confirm previous work showing that HAdV preferentially transforms cells of mesenchymal origin in rodents. More importantly, they demonstrate for the first time that human cells with stem cell characteristics can be completely transformed by HAdV oncogenes in tissue culture with high efficiency. Our findings strongly support the hypothesis that undifferentiated progenitor cells or cells with stem cell-like properties are highly susceptible targets for HAdV-mediated cell transformation and suggest that virus-associated tumors in humans may originate, at least in part, from infections of these cell types. We expect that primary hMSCs will replace the primary rodent cultures in HAdV viral transformation studies and are confident that these investigations will continue to uncover general principles of viral oncogenesis that can be extended to human DNA tumor viruses as well.
IMPORTANCE It is generally believed that transformation of primary human cells with HAdV-5 E1 oncogenes is very inefficient. However, a few cell lines have been successfully transformed with HAdV-5 E1A and E1B, indicating that there is a certain cell type which is susceptible to HAdV-mediated transformation. Interestingly, all those cell lines have been derived from human embryonic tissue, albeit the exact cell type is not known yet. We show for the first time the successful transformation of primary human mesenchymal stromal cells (hMSCs) by HAdV-5 E1A and E1B. Further, we show upon HAdV-5 E1A and E1B expression that these primary progenitor cells exhibit features of tumor cells and can no longer be differentiated into the adipogenic, chondrogenic, or osteogenic lineage. Hence, primary hMSCs represent a robust and novel model system to elucidate the underlying molecular mechanisms of adenovirus-mediated transformation of multipotent human progenitor cells.
INTRODUCTION
Human adenoviruses (HAdVs) are classified as DNA tumor viruses by virtue of their ability to cause tumors when inoculated into newborn rodents (1). Based on their oncogenicity in animals, the human subtypes can be divided into three classes: highly oncogenic adenoviruses, producing tumors at a high frequency within a few months; weakly oncogenic adenoviruses, inducing tumors inconsistently and after long incubation times; and nononcogenic adenoviruses (2). Despite the remarkable pattern of differential oncogenicity of HAdVs in animals, the members of all species tested so far are able to transform a variety of rodent cells in vitro with similar efficiencies.
In contrast, many attempts to transform primary human cells in culture with HAdVs have been unsuccessful, indicating that abortive infection, in which most of the early or all viral components have been synthesized but no infective virus is produced, is one of the factors associated with highly efficient transformation of nonpermissive rodent cells. However, transformation of human cells with subgenomic viral DNA fragments is extraordinarily inefficient compared to that in rodent cells, arguing that differences in permissivity to viral growth may not be the main determining factor in transformation efficiency (2–4). To date, only a few primary human cell types have been successfully transformed by HAdV-12, HAdV-5 DNA fragments, or HAdV-5 E1 oncogenes in culture, including human embryo kidney (HEK) cells (5), human embryonic lung (HEL) cells (6), human embryo retinoblasts (HER) (7), and amniotic fluid cells (AFC) (8). Among these, only HER and AFC can be reproducibly transformed, although less efficiently than rodent embryo or kidney cells. The molecular basis for the differences in transformation efficiencies between various human cell types is unknown (4). Previous work from Shaw et al. indicates that most of the transformed human cell lines that are derived from cultures of HEK and HER cells exhibit a pattern of intermediate filament expression similar to that seen in early differentiating neurons (9). Since HER cell cultures and, to a much lesser extent, HEK cell cultures contain cells of predominantly neuronal lineage, it has been proposed that human neuronal cells are a favored target for HAdV-mediated transformation. Whether transformed cells from transfections of AFC display a similar pattern of intermediate filament expression remains unknown.
In this report, we performed studies to identify target cells in the mixed AFC population that are susceptible to transformation by HAdV E1A/E1B oncogenes. We demonstrate that multipotent human mesenchymal stem cells (hMSCs) represent at least one cell type present in AFC that can be reproducibly transformed by HAdV-5 E1A/E1B as efficiently as primary baby rat kidney (BRK) cells. Moreover, we show that transformed hMSCs show phenotypic and genetic properties associated with a high grade of oncogenic transformation, including enhanced proliferation, anchorage-independent growth, and increased growth rates as well as numerical and structural chromosomal aberrations.
RESULTS
HAdV-5 E1A and E1B induce focus formation in primary multipotent hMSCs.To test whether HAdV-5 E1A/E1B oncogenes induce growth-promoting and transforming properties in primary human cells, we transduced low-passage-number bone marrow (BM)-derived hMSCs with HAdV-5 E1A and E1B. Freshly isolated pBRK cells were used as a positive control for E1A/E1B-mediated focus formation. Four weeks after transduction, cells were fixed and foci were stained with crystal violet (Fig. 1A). As expected, nontransduced hMSCs showed no focus formation, indicating that hMSCs in general are not prone to spontaneous immortalization. Also when primary hMSCs were transduced with empty vectors or E1A or E1B alone, we observed only a few, weakly dense foci, from which no stable cell lines could be established. In contrast, coexpression of HAdV-5 E1A and E1B oncogenes induced a statistically significant increase in dense and fast-growing foci (Fig. 1 and Tables 1 and 2), from which we generated several stable cell lines (hAB cells). Similar results were obtained using pBRK cells. Interestingly, the number of foci induced in hMSCs (Table 1) was almost equal to the number of foci obtained after transduction of pBRK cells with E1A and E1B (Table 2). These results demonstrate that primary hMSCs are highly susceptible to HAdV E1A/E1B-mediated focus formation and indicate that human cells with stem cell-like properties might be prime targets for HAdV-mediated cell transformation.
(A) Primary hMSCs were transduced with HAdV-5 E1A and/or E1B. iVLN2 and iBLB2 empty vectors were used as negative controls. Cells were fixed with crystal violet solution 4 weeks postransduction. Experiments were performed in triplicates. (B) Thirty micrograms of total cell lysate from subconfluent primary hMSCs and HEK293 and hAB cells was separated via 12% SDS-PAGE and analyzed for the indicated proteins. β-Actin was used as a loading control. (C and D) Cells were plated on glass coverslips, fixed with methanol 24 h postplating, and stained for E1A (C) or E1B-55K (D). Alexa Fluor 488-labeled secondary Ab was used to detect the primary Abs. 4′,6-Diamidino-2-phenylindole (DAPI) was used for nuclear staining. Overlays of single images (merge) are shown in subpanels c, g, k, o, s, and w (magnification, ×40). Bars, 25 μm.
Absolute focus numbers in transduced hMSCs
Absolute focus numbers in transduced pBRK cells
HAdV-transformed hMSCs present a deregulated protein expression pattern similar to that of other HAdV-transformed cell lines.Keeping in mind that HEK293 cells are of neuronal origin (9), whereas the novel established hAB cells are derived from hMSCs, we were interested in analyzing if coexpression of HAdV-5 E1A and E1B could induce changes in cellular protein expression similar to those observed in HEK293 cells, which constitutively express both E1A and E1B (7). In this line, it is known that adenoviral transformation is closely linked to interactions of E1A and E1B with cellular tumor suppressor proteins such as the retinoblastoma protein (pRb) or p53 and components of the DNA repair machinery like Mre11. E1A is known to phosphorylate pRb (10). Due to this, the E2F-1/RB complex that plays an important role in cell cycle progression cannot be formed (11). E2F-1 is released, and as a consequence, cells progress into S phase (12, 13). E1B-55K efficiently inhibits functions of the tumor suppressor p53 and DNA repair protein Mre11 (14, 15), which favors accumulation of mutations.
To further characterize the hAB cell lines, we analyzed the expression levels of the viral oncoproteins E1A and E1B as well as cellular proteins pRb, p53, Mre11, and E2F-1. HEK293 cells and primary hMSCs were used as references (Fig. 1B).
Western blot analysis revealed that all of the four hAB cell lines express E1A and E1B. E1A expression levels are similar in all hAB cell lines. However, compared to HEK293 cells only the faster-migrating E1A-12S isoform could be detected. It seems that during the propagation, a selection toward the less-toxic E1A-12S (243R) isoform took place. Furthermore, E1B-55K steady-state concentrations differ between the analyzed hAB cell lines, which might be due to different numbers of integrated E1B sequences. More important, all hAB cell lines as well as HEK293 cells show distinctly increased expression levels of pRb, Mre11, p53, and E2F-1 compared to the parental hMSCs.
This shows that, even though the origin of the cell lines is different from HEK293 cells, the transformation process triggered in hMSCs by adenoviral E1 oncogenes seems to induce a similarly deregulated protein expression. Increased pRb and E2F-1 levels might appear due to higher proliferation rates of hAB and HEK293 cells than of hMSCs. Furthermore, enhanced p53 levels are a result of E1A, which is known to induce and stabilize p53 (16). Moreover, increased Mre11 concentrations might indicate that there is an Mre11-mediated induction of the DNA damage response (DDR), due to the E1A expression and resulting onset of unscheduled DNA replication.
E1A and E1B subcellular localization in transformed hMSCs.To characterize more closely the transformed hMSCs, E1A and E1B expression levels were compared between the different hAB cell lines, primary hMSCs, and HEK293 cells (Fig. 1B). E1A and E1B protein levels varied among the transformed cell lines. Significantly, hAB3 to hAB5 cells expressed lower E1B-55K levels than HEK293 cells. Normalization to β-actin revealed that E1B-55K levels in hAB3 to hAB5 cells were reduced up to 75% whereas hAB6 cells expressed E1B-55K almost as highly as HEK293 cells (data not shown). The molecular basis for this effect is unknown but may be related to differences in E1A expression levels in these cell lines. In addition, HEK293 cells expressed both E1A forms (E1A-12S and E1A-13S), while all hAB cells produced only the smaller E1A-12S protein, although the complete E1A region, encoding both viral polypeptides, was transduced.
Consistent with previous reports (17), we observed that E1A localized diffusely in the nucleus in HEK293 and hAB cells when we analyzed the subcellular localization of the adenoviral proteins (Fig. 1C). Furthermore, E1B-55K was found concentrated in a large, partly spindle-shaped perinuclear body (Fig. 1D), which is typically observed in all human and rodent cells that are transformed by species C HAdVs (18). Altogether, these data demonstrate that all hAB cell lines exhibit characteristic features of HAdV-transformed rodent and human cells, including stable expression of viral oncogenes and accumulation of p53 and pRb as well as deposition of E1B-55K into large cytoplasmic aggregates located in close proximity to the nuclear membrane.
HAdV-5 E1A and E1B expression in hAB cells elicits enhanced cell growth and proliferation rates.Previously, we and others have demonstrated that transformed BRK cells stably expressing adenoviral oncogenes display distinctly increased proliferation rates and grow to higher densities (19). To investigate if the HAdV-5 E1 oncogenes induce similar changes in hMSCs, we analyzed the growth rate of hAB cells.
Equal amounts of the different hAB cell lines were seeded, and cells were counted every 72 h over a period of 9 days. Our results showed that all four hAB cell lines displayed increased growth rates compared to primary hMSCs (Fig. 2A). Interestingly, the strongest effect on cell proliferation and growth density was observed in hAB3 and hAB4 cells, which express higher levels of E1A and E1B. Already at 3 days postplating, these cells showed significantly higher proliferation rates than the other cells and reached approximately 50- and 100-times-higher growth densities than the parental hMSCs at the end of the experiment (Fig. 2A).
(A) Cells (1 × 104/well) were plated on six-well dishes, and viable cells were counted every 72 h. (B) Cells (5 × 104/well) were plated on 12-well dishes. MTT reagent was added 2 h before measurement. Results are expressed as ratios: values at 24 h were subtracted from the ones at 72 h and then normalized to parental hMSCs. Both graphs show the mean value and standard deviation from three independent experiments. The asterisks represent the t test P value (****, P < 0.0001). (C) TRAP assay was used to analyze telomerase activity by detecting a ladder of products with 6-base increments starting at 50 nucleotides: 50, 56, 62, 68, etc. A representative ethidium bromide-stained polyacrylamide gel is shown. Positive controls for telomerase activity after incubation of a 500- or 1,000-cell-equivalent probe (e.p.) with an artificial telomere repeat template (lanes 1 and 17, respectively). These ladder controls show bands starting at 50 bp with a 6-bp increasing difference. A semicompetitive 36-bp internal control (int. ctrl) was included in each sample. Two quantification controls (lanes 3 and 4) as well as HEK293 cell probes (lanes 7 and 8) were included. Results for hMSCs (lanes 5 and 6) and HEK293 (lanes 7 and 8) and hAB (lanes 9 to 16) cell lines are shown. Asterisks (*) mark heat-inactivated samples. The heat-inactivated control samples were incubated at 85°C for 10 min prior to TRAP assay. The arrowheads indicate primer-dimer artifacts, which are clearly distinguishable from the laddering pattern resulting from genuine telomerase activity. (D) Representative images of the soft agar colony formation assay are shown for each hAB cell line (c and d) and hMSC par (a). HEK293 cells were included as a positive control (b). Images were taken on an inverted light microscope (Leica DMIL/DFC320) with a phase-contrast filter using an ×10 magnification. hMSC par, parental hMSCs.
These findings are supported by a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) cell proliferation assay. In this assay, the metabolic activity of the different hAB cell lines as well as the parental hMSCs was measured over a period of 72 h. The metabolic activity of all hAB cells was distinctly increased compared to the parental hMSCs, which is concordant with enhanced cell growth (Fig. 2B).
Taken together, these results demonstrate that the coexpression of HAdV-5 E1A and E1B induces alterations in the hMSC metabolism associated with the transformed phenotype, allowing these cells to grow more rapidly and to a higher density.
HAdV-5 E1A/E1B-expressing hMSCs show increased telomerase activity.One hallmark of immortalized cells is the onset of telomere maintenance mechanisms in order to prevent the degradation of chromosome ends. One vital difference between primary rodent and human cells is that rodent cells constitutively express telomerase, making them more susceptible to transformation processes (3).
However, telomerase activity in hMSCs is not detectable (20) or only to a very small extent (21). In order to investigate whether immortalized hAB cell lines regained telomerase activity, we performed a telomere repeat-amplification protocol (TRAP) assay (Fig. 2C).
Consistent with previous reports (20, 21), primary hMSCs showed no telomerase activity (Fig. 2C, lanes 5 and 6). However, all hAB cell lines, as well as HEK293 cells, displayed strong telomerase activity, indicating that this feature was acquired upon HAdV-5 E1A/E1B coexpression, supporting their transformed phenotype.
E1A- and E1B-positive hAB clones show anchorage-independent growth.Due to the fact that all hAB cell lines exhibit strong telomerase activity, we were interested in additional hallmarks of cancer cells that hAB cells might have acquired.
Another key aspect of transformed cells is the ability to grow in an anchorage-independent manner.
To analyze if the hAB cell lines developed anchorage-independent growth properties, a soft agar colony formation assay was performed and hAB cells were compared to HEK293 cells and parental hMSCs. Cells were cultivated and analyzed 4 weeks postplating (Fig. 2D). As expected, primary hMSCs did not form colonies in soft agar, indicating that they are anchorage dependent (Fig. 2D, a). In contrast, HEK293 cells formed symmetrical spheroidal colonies, which are an indicator for anchorage-independent growth (Fig. 2D, b). Interestingly, all hAB cell lines could form colonies in soft agar (Fig. 2D, c to f), showing the acquisition of another important feature of transformed cells. However, compared to high-passage-number HEK293 cells, colonies formed by hAB cells were fewer in number (hAB3, 40%; hAB4, 15%; hAB5, 20%; hAB6, 10%), irregular in shape, and heterogeneous in size, indicating that they differ phenotypically even though they were derived from single-cell clones (Fig. 2D, e and f).
hAB clones are genetically unstable and have a complex karyotype.Genomic instability is characteristic for transformed cells. Chromosome rearrangements, additions or deletions, and mutations are some of the consequences of this instability. Knowing that HAdVs are able to induce DNA damage (22) and due to observations obtained from the soft agar assay, where colonies formed by hAB cells were heterogeneous in shape and size, we decided to investigate if the diverse phenotypes corresponded to complex genotypes by multicolor fluorescent in situ hybridization (M-FISH) analysis.
At least 10 metaphases per cell line were analyzed. Representative images of parental hMSC and hAB cell lines are shown in Fig. 3.
(A) M-FISH analysis of parental hMSCs (P4) from a healthy female donor. (B to E) M-FISH analysis of hAB cell lines (P15) with multiple chromosomal alterations. Representative spectral images of an hMSC diploid metaphase (hMSC par) and tetraploid metaphases from hAB cell lines (hAB3 to -6) are shown. The detailed genotype of the corresponding metaphase image and clonal changes within a series of analyzed metaphases are listed in Table 3.
As expected, the parental hMSCs show a normal karyotype of a female individual with a set of 23 chromosome pairs without genetic rearrangements (Fig. 3A). Interestingly, persistent expression of HAdV-5 E1A/E1B oncoproteins led to highly unstable karyotypes among the hAB cells characterized by genomic aberrations, including numerical and structural chromosomal alterations as well as formation of newly arranged marker chromosomes (Fig. 3B to E). Interestingly, some of the various chromosomal changes were detected in each analyzed metaphase of the corresponding hAB cell line (Table 3). This indicates that all analyzed metaphases within a set were derived from the same origin. However, due to the coexpression of HAdV-5 E1A and E1B, which interfere with the cellular DNA damage response, progeny cells can randomly acquire more and more chromosomal mutations, which presumably is the cause of genomic instability.
Summary of karyotypes and clonal changes observed by M-FISHa
Transformed hAB cells have lost their differentiation potential.Given that hMSCs are multipotent progenitor cells, which can be differentiated into adipogenic, chondrogenic, and osteogenic lineages (23), we tested if our novel generated hAB cell lines were still multipotent. For this, we cultivated the different hAB cell lines as well as primary hMSCs and induced differentiation into adipocytes, chondrocytes, or osteoblasts.
Our results clearly showed that the parental hMSCs could be successfully differentiated into the adipogenic (Fig. 4A), chondrogenic (Fig. 4B), or osteogenic lineage (Fig. 4C), whereas the novel generated hAB cell lines have completely lost their differentiation potential into any of the aforementioned lineages, which reflects another acquired feature of tumor cells.
Representative images from parental hMSCs and hAB cells after induction of differentiation. All cell lines were grown in either normal medium (control) or special differentiation medium (induced). (A) Staining of adipocyte-like cells with Sudan red (b). (B) Staining of chondrocyte-like cells with alcian blue (b). (C) Optical dense calcium precipitates indicate differentiation into the osteogenic lineage (b). Light microscopic images were taken at an ×20 magnification (Leica DMIL/DFC320).
DISCUSSION
Successful transformation of primary hMSCs by coexpression of HAdV-5 E1A/B.Several types of human cancers are closely associated with viral infections (24). For adenoviral E1 oncogenes, it has been shown that they can transform human cells in vitro (5, 7, 8). However, so far this transformation process is not well understood and occurs only rarely. Moreover, it is still unclear what kinds of cells are prime targets for adenoviral transformation. There are some indications that putative target cells originate from neuronal cells. Shaw and coworkers found that HEK293 cells, which constitutively express E1A and E1B genes, are positive for neural markers (9). Others state that target cells for adenoviral transformation exhibit both mesenchymal and neural characteristics (25).
Our results support the hypothesis that target cells are of mesenchymal origin.
However, it seems to be even more important that HAdV E1 oncogene-transformable cells are undifferentiated cells (stem cells) or at least in an early differentiation stage. It has been reported that newborn hamsters developed undifferentiated sarcomas after injection of adenovirus (1). In this context, our findings support mesenchymal progenitor cells as putative targets for adenoviral transformation, as sarcomas are cancers of mesenchymal origin.
Furthermore, human amniotic fluid cells can also be transformed by HAdV-5 E1A and E1B sequences (8). However, AFC are a heterogeneous cell population composed of a large percentage of hMSCs as well as stem cells and human amniotic fluid epithelial cells (26). This is in line with our findings pointing to multipotent hMSCs as putative targets for adenoviral transformation. However, while previous studies using primary human AFC have shown only one or two transformation events per 1 × 105 cells (8), we have obtained a transformation rate of one event per 3 × 102 cells when using purified hMSCs.
Another important aspect to consider is the usage of lentiviral vectors for efficient HAdV-5 oncogene delivery. There was no dense focus formation in hMSC cells transduced by either E1A or E1B and the corresponding empty lentiviral vectors, which implies that integration mutagenesis does not cooperate with E1A or E1B functions. However, lentiviral integration might support the transforming functions of the adenoviral oncogenes by ensuring the stable expression of E1A and E1B over time.
Further, the usage of lentiviral vectors for adenoviral E1 oncogene delivery has several advantages over plasmid transfection. For instance, the transduction efficiency is higher and less toxic for primary human cells than plasmid transfection (27). This might be an explanation for inefficient transformation of human cells reported in past reports.
However, attempts to transform differentiated cells by HAdV-5 E1A and E1B after lentiviral delivery have been unsuccessful (data not shown), which implies that there are additional factors making hMSCs susceptible to HAdV-5 E1A/E1B-mediated transformation.
This strongly supports the hypothesis that multipotent hMSCs are susceptible to efficient HAdV-5 E1A/E1B oncogene-mediated transformation and indicates that stem cell-like characteristics represent an important prerequisite for transformation processes and may open up new prospects by combining virus-mediated oncogenesis and stem cell biology. Moreover, our results support the general hypothesis in which stem cells, or at least stem cell-like progenitor cells, might be susceptible not only to HAdV but also to other virus-mediated transformation processes.
E1A/E1B-expressing hMSCs display distinct hallmarks of transformed cells.We thoroughly studied the phenotype of hMSCs that had been transformed by HAdV-5 E1A/B oncogenes. In detail, we analyzed the cells regarding E1A and E1B expression, telomerase activity, anchorage-independent growth, steady-state expression levels of tumor suppressor genes, cell differentiation, and genetic stability.
Interestingly, our novel hAB cell lines exhibit enhanced cell proliferation and altered growth properties compared to primary hMSCs. Increased proliferation rates are, among others, an acquired capability of cancer cells, when those become insensitive to antigrowth signals. Many antiproliferative signals are funneled through pRb and its two relatives, p107 and p130 (28). E1A interacts with pRb, leading to the liberation of E2F transcription factors, and thus allows unscheduled cell proliferation, rendering cells insensitive to antigrowth factors (12). Further, pRb blocks the E2F function, which controls the expression of genes essential for S-phase progression (29). One consequence of the E1A-induced cell proliferation is the stabilization of p53 and the induction of apoptosis, which is efficiently blocked by the E1B-mediated inhibition of p53 (30). Therefore, evading apoptosis displays an additional cancer cell capability that hAB cells have acquired.
Unlimited growth is another substantive feature of transformed cells. Preventing telomere shortening is an important prerequisite for cell immortalization. Thus, mechanisms of telomere maintenance are evident in all malignant cells (31). Transformation of somatic cells has been associated with telomerase activation and preservation of telomeres. The central mechanism of telomerase activation seems to be a change in the transcriptional control of the human telomerase reverse transcriptase gene (hTERT), encoding the catalytic subunit of telomerase (32).
We could show that primary hMSCs exhibit no telomerase activity in the TRAP assay. This is consistent with findings from others in which telomerase activity was absent or only basal activity could be detected (20, 21).
Interestingly, artificial overexpression of hTERT extends the life span of hMSCs by maintaining a normal karyotype and their multilineage differentiation potential (33).
One explanation for the onset of telomerase in hAB cells might be that hTERT expression has been directly or indirectly targeted by E1A or E1B. It has been shown that p53 is also an important suppressor of hTERT transcription (34). However, E1B-55K is able to efficiently counteract p53 activity (35). Therefore, E1B-mediated inhibition of p53 could have triggered hTERT expression in hAB cells. Further telomere-independent effects of telomerase include enhancement of cell proliferation and resistance to apoptosis (36). These functions might have been additional driving forces together with the growth-promoting properties of E1A and E1B for the successful propagation of the hAB cell lines.
In line with this model, it appears conclusive that transformation of telomerase-negative primary human cells is believed to be extraordinarily ineffective (2). It is conceivable that telomerase activity and stem cell-like self-renewal capabilities render primary hMSCs susceptible to HAdV E1 oncogene-mediated transformation.
Strikingly, all hAB cell lines exhibited strong telomerase activity. Even though the exact mechanism is not clear so far, this finding suggests that the growth-stimulating functions of E1A together with E1B have the potential to keep or to upregulate telomerase activity during transformation of primary hMSCs. Besides this, the hAB cell lines displayed further features of malignant transformation, including loss of contact inhibition as well as anchorage-independent growth, which is considered one of the key features of transformation in vitro and aggressive tumors in vivo (37). Interestingly, there was broad diversity in three-dimensional (3D) colony formation among the hAB cell lines, which might result through different proliferation capabilities of genetically instable cells.
Indeed, one of the most impressive findings was the high grade of genomic rearrangements, which displayed an enormous variety on the chromosomal level, when hAB cells were subjected to M-FISH analysis. A possible reason for this could be the sustained expression of HAdV-5 E1 oncogenes. For instance, E1B-55K is known to interfere with the cellular DNA damage response by inactivating the MRN complex (14), which contributes to the vast genomic rearrangements.
Given the fact that hAB cells are genetically unstable and exhibit a huge number of chromosomal deletions, insertions, and duplications as well as small supernumerary marker chromosomes (38, 39), it seems probable that a prolonged period of cultivation might positively select for the fastest-growing and potentially more malignant phenotypes.
Therefore, we analyzed if the hAB cells still exhibit characteristics of primary hMSCs. hAB cell lines were subjected to differentiation into the adipogenic, chondrogenic, or osteogenic lineages. Interestingly, none of the novel hAB cell lines could be induced to differentiate into adipocytes, osteoblasts, or chondrocytes, whereas the parental hMSCs were able to differentiate into all three lineages. This is in clear contrast to hMSC-derived cell lines that have been immortalized by hTERT overexpression alone or in combination with human papillomavirus 16 E6/E7 oncogenes (HPV16 E6/E7), as those maintained their differentiation potential (40). Furthermore, simian virus 40 (SV40) Tag-immortalized hMSCs also maintained their mesodermal trilineage differentiation potential (41).
However, our results can be at least partially explained by former findings in which E1A was shown to contribute to epigenetic reprogramming of the host cell during infections, which is crucial for an efficient virus replication but is also important for transformation processes. For instance, E1A activates genes that are necessary for cell cycle progression from G0 to the S phase. Further, it represses numerous genes involved in differentiation, through its interaction with the pocket proteins (pRb, p130, and p107) or with the p300/CBP complex (42, 43). These findings point toward an E1A-mediated block of differentiation in hAB cells. Our assumption is supported by earlier studies, where muscle-specific genes were suppressed by HAdV-5 E1A functions, which induced dedifferentiation of muscle cells (44). Further, E1A is able to repress differentiation in certain cell lines (45). Even more interesting in this context, Cao and coworkers showed that HAdV-5 E1A inhibits differentiation of rodent preadipocytes by its interaction with pRb (46).
In conclusion, we have shown for the first time that we are able to transform primary multipotent human mesenchymal stromal cells with HAdV-5 E1A and E1B oncogenes.
The established E1A- and E1B-expressing hAB cells acquired and exhibited several hallmarks of malignant transformation, including loss of their differentiation potential, reactivation of telomerase, anchorage-independent growth, and a high grade of genetic instability. Furthermore, our data imply that adenoviral E1 oncogenes are highly suitable to study adenovirus-mediated transformation on a molecular level not only in the primary rodent system but also in primary human cells. Our data suggest that hMSCs are prone to adenoviral transformation and represent a reservoir in vivo that might give rise to adenovirus-mediated oncogenesis. Moreover, they emphasize that primary human cells with progenitor or stem cell characteristics might represent prime targets for HAdV-5 E1 oncogene-mediated or virus-induced transformation in general. These studies provide a new model to study the mechanisms driving virus-mediated transformation in humans.
MATERIALS AND METHODS
Primary cells and cell lines.Human mesenchymal stromal cells (hMSCs) were obtained from bone marrow (BM) donated by healthy donors after informed consent and according to the hospital's guidelines approved by the Hamburg Ethics Committee. hMSCs were generated and expanded as described by Lange et al. (47). Briefly, 150 μl unprocessed BM was seeded into T75 cell culture flasks (Sarstedt) in expansion medium consisting of DMEM-LG (Dulbecco's modified Eagle's medium-low glucose, Gibco) plus 2 mM GlutaMAX (Gibco) plus 10% preselected fetal calf serum (FCS) (BioWhittaker) and incubated at 37°C and 5% CO2. After 2 to 3 days, nonadherent cells were washed off. Cells were fed twice a week with fresh medium until they reached 90 to 95% confluence. This time point was defined as starting point P0. Afterward, cells were detached with 0.05% Trypsin-EDTA (Gibco), and 500 cells were seeded per cm2 and designated P1.
Primary baby rat kidney (pBRK) cells were obtained from 3- to 5-day-old CD IGS rats (Charles River) as described previously (19). Briefly, organs were incubated with 1 mg/ml collagenase-dispase (Roche) at 37°C for 3 h and single cells were seeded in DMEM supplemented with 10% FCS.
HEK293 human embryonic kidney cells (ATCC CRL-1573) (5) were grown in DMEM supplemented with 10% FCS, 100 U penicillin, and 100 μg streptomycin per ml in a 5% CO2 atmosphere at 37°C.
Cell line authentication.hMSCs and transformed hMSC-derived HAdV-5 E1A- and E1B-expressing cells (hAB cell lines) have been authenticated by the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany) using nonaplex PCR to generate DNA profiles of eight highly polymorphic loci of short tandem repeats (STRs). Additionally, cells have been tested to be free of mouse, rat, and hamster DNA. The STR profile of the parental hMSCs (hMSC par) does not match any STR reference profile of the international STR database and is therefore unique. The STR profiles of the hAB cell lines correspond to the hMSCs from which those are derived and have also been tested to be free of mouse, rat, and hamster DNA.
Generation, production, and titration of lentiviral vectors.Lentiviral gene ontology (LeGO) vectors (48, 49) were used for delivery of adenoviral sequences. PCR-amplified genomic HAdV-5 E1A (GenBank accession no. AY339865 , nucleotides [nt] 560 to 1545) and E1B (GenBank accession no. AY339865 , nt 1714 to 3630) regions were ligated into the respective lentiviral vector. E1A was cloned into the LeGO-iVLN2 backbone, which constitutively expresses the gene of interest together with an internal ribosome entry site (IRES)-linked Venus-neomycin resistance fusion protein under the control of the spleen focus-forming virus (SFFV) promoter. E1B was cloned into the LeGO-iBLB2 vector backbone, which is similar to the abovementioned but carries blue fluorescent protein (BFP)-blasticidin S resistance instead of the Venus-neomycin resistance sequence.
Lentiviral particles were produced and titrated according to the method of Weber et al. (48). Briefly, particles were pseudotyped with vesicular stomatitis virus G protein (VSV-G). Titration was performed on HEK293T cells, and titers were determined using a FACSCanto-II cytometer (BD Bioscience).
Transduction of primary mammalian cells.hMSC or pBRK cells were seeded on 12-well plates (Sarstedt) 12 h before transduction. Adherent cells were transduced at multiplicities of infection (MOIs) of 0.3 particles/cell as described previously (48). Cells were cultivated for 3 to 4 weeks in DMEM-LG (Dulbecco's modified Eagle's medium-low glucose; Gibco) supplemented with 2 mM GlutaMAX (Gibco), 10% preselected FCS (BioWhittaker), 200 μg/ml G418 (Calbiochem), and 50 μg/ml blasticidin S (InvivoGen). Antibiotics were added to select for E1A- and E1B-positive cells. Medium was changed every 4 days. Single foci were isolated and subsequently expanded into permanent monoclonal cell lines. Alternatively, foci were stained with crystal violet (1% in 25% methanol) and counted.
Antibodies.Primary antibodies specific for HAdV-5 proteins included E1A mouse monoclonal antibody (MAb) M73 (50) and E1B-55K mouse MAb 2A6 (51). Primary antibodies against specific cellular and ectopically expressed proteins included E2F-1 (sc-193; Santa Cruz), mouse MAb pRb (9309; Cell Signaling Technology), Mre11 rabbit polyclonal antibody (PAb) (pNB 100-142; Novus Biologicals), β-actin mouse MAb AC-15 (A5441; Sigma-Aldrich), and p53 mouse MAb DO-1 (sc-126; Santa Cruz). Secondary antibodies conjugated to horseradish peroxidase (HRP) against mouse or rabbit IgG have been obtained from Dianova. Alexa Fluor 488-labeled secondary antibodies against mouse IgG have been obtained from Life Technologies (A11017).
Western blotting.Cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol [DTT], 0.1% sodium dodecyl sulfate, 1% Nonidet P-40, 0.1% Triton X-100, 0.5% sodium deoxycholate) containing freshly added 1% (vol/vol) phenylmethylsulfonyl fluoride (PMSF), 0.1% (vol/vol) aprotinin, 1 μg/ml leupeptin, 1 μg/ml pepstatin, and 1 mM DTT. Protein concentrations were measured using the Bio-Rad protein assay. Equal amounts of total protein were separated by SDS-PAGE and transferred to nitrocellulose membranes by Western blotting. Membranes were blocked for 2 h in phosphate-buffered saline (PBS) containing 0.1% Tween 20 and 5% nonfat dry milk powder. Afterward, membranes were incubated for 2 h in PBS containing 0.1% Tween 20 and the appropriate primary antibody. Proteins were visualized by secondary HRP-conjugated antibodies followed by enhanced chemiluminescence (ECL), as recommended by the manufacturer (Pierce) on medical X-ray films (CEA RP). Autoradiograms were scanned and cropped using Adobe Photoshop CS6 and assembled with Adobe Illustrator CS6.
Indirect immunofluorescence.Cells were grown on glass coverslips as described previously (52). Cells were fixed in methanol at −20°C for 15 min and permeabilized in PBS plus 0.5% Triton X-100 at room temperature for 30 min. Coverslips were blocked for 1 h. Afterward, the blocking solution was discarded and the cells were incubated for 1 h with respective primary antibodies. Samples were washed three times in PBS plus 0.1% Tween 20 and subsequently incubated with the corresponding Alexa Fluor 488-conjugated secondary antibodies for 30 min. After an additional washing step, samples were covered with antifade solution (Energene) and digital images were acquired on a DMI6000 fluorescence microscope (Leica) with a charge-coupled device (CCD) camera. Images were cropped using Adobe Photoshop CS6 and assembled with Adobe Illustrator CS6.
MTT assay.The MTT [3-(4,5-dimethylthiazolyl-2-yl)-2,5-diphenyltetrazolium bromide] (Sigma-Aldrich; St. Louis, MO) assay was used to evaluate cell viability. A total of 5 × 104 cells/well was grown overnight on 12-well plates. MTT solution was added to each well to a final concentration of 500 μg/ml and incubated for 2 h at 37°C. Afterward, medium was removed and formazan crystals were dissolved in 200 μl of N-propyl alcohol containing 0.1 N HCl. One hundred microliters of this solution was transferred to a flat-bottom 96-well plate, and absorbance was measured at 540 nm using a microplate reader (SynergyMX; BioTek). All values were determined in triplicates, and data were analyzed using Prism5 software (GraphPad).
Growth curve.Cells at 1 × 104 cells/well were initially seeded onto 6-well plates in DMEM supplemented with 5% FCS plus 100 U/ml penicillin and 100 μM streptomycin mix. Medium was replaced every 48 h. Viable cells were trypsinized at indicated time points postplating and counted in a Neubauer hemocytometer. Cell viability was determined with trypan blue exclusion dye.
Soft agar colony formation assay.Thirty-five-millimeter dishes were coated with cultivation medium containing 0.5% agar. A total of 5 × 103 cells suspended in DMEM containing 0.3% agar and 10% fetal bovine serum (FBS) was seeded onto the first layer. Each cell type was plated in triplicate. Cells were incubated under standard cell culture conditions (37°C, 5% CO2, 95% H2O). Once a week, 0.7 ml of DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μM streptomycin mix was added on top of the semisolid medium to feed the cells. Cell colonies were imaged at a 100-fold magnification on an inverse light microscope (DMIL/DFC320; Leica).
TRAP assay.Telomerase activity from HAdV-5 E1A/E1B-positive hAB cells was determined with the telomere repeat-amplification protocol (TRAP) (53) using the TRAPeze telomerase detection kit (Merck Millipore). Detection of telomerase activity using the TRAP assay in cultured cells involves the addition of TTAGGG repeats by telomerase to a substrate oligonucleotide (TS) and the subsequent PCR amplification of these extension products with both the forward (TS) and reverse (CX) primers, generating a ladder of products with 6-base increments starting at 50 nucleotides: 50, 56, 62, 68, etc. The TRAPeze telomerase detection kit (Merck Millipore) was used as recommended by the manufacturer's instructions. Briefly, cell pellets were stored at −80°C until lysis was performed. Approximately 106 cells were harvested and lysed in 200 μl of 1× CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate} lysis buffer (Tris-HCl, 10 mM, pH 7.5, 1 mM MgCl2, 1 mM EGTA, 0.1 mM benzamidine, 5 mM β-mercaptoethanol, 0.5% CHAPS, 10% glycerol) and incubated on ice for 30 min. Cell debris was precipitated for 20 min at 12,000 × g at 4°C. The supernatant was flash-frozen and stored at −80°C. For the PCR, 2 μl of extract (corresponding to 500 cells) was combined with the 48-μl reaction mixture supplied with the kit and 2 units of Taq DNA polymerase (Fermentas). The heat-inactivated control samples were incubated at 85°C for 10 min prior to TRAP assay. After incubation for 30 min at 30°C (telomere elongation step), samples were heated to 95°C for 2 min to inactivate telomerase; afterward, the samples were subjected to 33 PCR cycles of 94°C for 15 s and 59°C for 30 s (FlexCycler; Analytik Jena). The PCR products were separated by electrophoresis on nondenaturing 12.5% polyacrylamide gels. Gels have been stained with ethidium bromide. The typical 6-bp DNA incremental ladder of telomerase products was detected using the G:Box gel documentation system (Syngene). An important feature of the TRAPeze gel-based telomerase detection kit is the inclusion of an internal control to monitor PCR inhibition in every lane, which is included in the PCR primer mix. The kit also provides a telomerase quantitation control, which is an oligonucleotide with a sequence identical to the TS primer extended with 8 telomeric repeats, AG(GGTTAG)7. This control serves as a standard for estimating the amount of TS primers with telomeric repeats extended by telomerase in a given extract.
Multicolor FISH assay.Cells were exposed to colchicine for 4 h. Subsequently, cells were treated with 0.075 M KCl for 20 min, fixed with a methanol-acetic acid solution (3:1), and subjected to multicolor karyotyping by FISH (M-FISH). The M-FISH assay was performed using the 24XCyte color kit for human chromosomes (MetaSystems) according to the supplier's recommendations. Briefly, metaphases were hybridized with chromosome probes for all chromosomes simultaneously. Each probe was labeled with one of five different fluorochromes (diethylaminocoumarin [DEAC], fluorescein isothiocyanate [FITC], Spectrum Orange, Texas Red, and Cy5) or a unique combination of them. DNA was counterstained with 4′,6-diamidino-2-phenylindole. After hybridization, grayscale images of the fluorochromes were acquired using an epifluorescence microscope (Carl Zeiss) equipped with a high-resolution cooled CCD camera (Photometrics). A 24-pseudocolor image was built up by overlay of the grayscale images and analysis by the MetaSystems Isis software package (MetaSystems). M-FISH is based on the individual fluorescence labeling of each chromosome. Here, numerical and structural chromosomal aberrations, even those which are cryptic or not further specified by conventional cytogenetic techniques, can be precisely detected and described. Five to 35 well-conserved and complete abnormal metaphases were analyzed at approximately 350 to 450 band levels. Karyotypes were described according to the International System for Human Cytogenetic Nomenclature (ISCN 2005) and revised using the CyDAS online analysis site (54).
Differentiation assay.Assays for differentiation to adipo-, chondro-, and osteogenic lineages were carried out as described previously (55). After induction, fat accumulations in adipocytes were visualized with Sudan red, proteoglycans were secreted from chondroblasts with alcian blue, and calcium deposits in osteoblasts were stained with silver nitrate. To ensure direct comparability, the induction time for adipo-, chondro-, and osteogenic differentiations was kept constant for all cell preparations.
ACKNOWLEDGMENTS
The Heinrich Pette Institute is supported by the Freie und Hansestadt Hamburg and the Bundesministerium für Gesundheit. This work was supported by grants from the Wilhelm Sander Stiftung and the Deutsche Forschungsgemeinschaft (Do 343/7-1).
We declare no conflict of interest.
FOOTNOTES
- Received 2 September 2016.
- Accepted 15 October 2016.
- Accepted manuscript posted online 19 October 2016.
- Copyright © 2016 American Society for Microbiology.
