|Cancer||Cell death||Cell cycle||Cytoskeleton||Exo/endocytosis||Differentiation||Division||Organelles||Signalling||Stem cells||Trafficking|
Functional heterogeneity of non-small lung adenocarcinoma cell sub-populations
Iga Bechyne, Katarzyna Szpak, Zbigniew Madeja and Jarosław Czyż1
Department of Cell Biology, Faculty of Biochemistry, Biophysics and Biotechnology, Jagiellonian University, Cracow, Poland
The morphological and functional heterogeneity of solid tumour cells can be observed in cancer cell lines cultured in vitro. We have combined analyses of microclones developed from single cells with micropore transmigration assays to demonstrate the co-existence of cellular subsets differing in morphology and motile activity, as well as Cx43 (connexin 43) and N-cadherin expression within lung carcinoma A549 populations. ‘Fibroblastoid’ cells, characterized by high motility, polarized morphology and plasmalemmal localization of Cx43, displayed the strongest aptitude for transmigration through narrow obstacles. Due to high mitotic activity, they maintain the whole population but can also give rise to a sub-population of quiescent and immobile ‘epithelioid’ cells. Our observations indicate that phenotypic transitions between the fibroblastoid and epithelioid phenotype account for the heterogeneity of metastable A549 cell populations.
Key words: connexin 43 (Cx43), invasion, lung, lung cancer, metastability
Abbreviations: Cx43, connexin 43, EMT, epithelial–mesenchymal transition, NA, numerical aperture; TLCD, total length of cell displacement
1To whom correspondence should be addressed (email email@example.com).
In principle, the development of solid tumours depends on the accumulation of epigenetic modifications within the progeny of a single cell (Visvader, 2011). Erroneous exchange of signals between cancer cells and their microenvironment (Sottoriva et al., 2010a) participates in the formation of heterogeneous, functionally specialized cellular phenotypes within a developing tumour (Sottoriva et al., 2010b). Among these appear sub-populations of invasive cells that are capable of migrating over long distances and are predestined to cross tissue barriers (Gupta and Massague, 2006; Langley and Fidler, 2007). Cellular subsets differing in basic properties crucial for cancer development were also found within in vitro cultured cancer cell populations (Langley and Fidler, 2007). This heterogeneity is not necessarily due to the co-existence of independent cell populations, but can result from their metastability, implying the ability of cancer cell sub-populations to convert into one another.
EMT (epithelial–mesenchymal transition; Thiery et al., 2009; Savagner, 2010) is an example of phenotypic transition within cancer cell populations. It is an epigenetic process leading to the formation of highly mobile ‘mesenchymal’ or ‘fibroblastoid’ cells from non-motile ‘epithelial’ cell populations. Cell motility is a prerequisite for cancer invasion and formation of metastases; therefore EMT seems crucial for cancer progression in vivo (Berx et al., 2007). Mesenchymal cells representing an invasive cancer cell sub-population may comprise a subset of secondary tumour stem cells, i.e. the cells from which metastases originate. However, epigenetic memory of mesenchymal cells resulting in their metastability may prompt transitions towards the ‘epithelioid’ phenotype important for the maintenance of secondary tumours. The so-called MET (mesenchymal–epithelial transition) reportedly participates in the formation of epithelioid cells which form an envelope stabilizing micrometastases and facilitating the formation of macroscopic tumours (Hugo et al., 2007). Phenotypic transitions between metastable states of cancer cells can occur spontaneously due to stochastic fluctuations of transcriptional regulators, or be induced in response to external factors (Graf and Stadtfeld, 2008). We have used an experimental approach based on combined microclone tests and micropore transmigration assays to identify the origin of heterogeneous sub-populations within the lung adenocarcinoma A549 cell line.
2. Materials and methods
2.1. A549 cell cultures
Human lung carcinoma A549 cells (ECACC 86012804) were cultivated in RPMI 1640 medium (Sigma) supplemented with a 10% FCS (fetal calf serum; Gibco) and MycoGONE mycoplasma antibiotics cocktail (AMSBio) in a humidified atmosphere with 5% CO2 in air at 37°C. In micropore transmigration experiments, A549 cells were seeded into chambers containing microporous membranes (Corning; pore diameter 8 μm; membrane diameter 5 mm) at 300 cells per mm2 and allowed to transmigrate for the next 48 h, i.e. sufficient time for the first cells (<1% of total cell number) to precipitate on to the well bottom. Afterwards, their progeny (A549T1 population) was used for end-point experiments or was further propagated. In short, A549T1 cells were trypsinized and seeded to the secondary dishes at 104 cells per cm2, cultivated for 96 h and either transferred to the next culture dish (to obtain A549T2 and, prospectively, T3 and T4 populations) or used for end-point experiments as the A549T2 population. For ‘microclone’ assays, cells were seeded at 500 cells per cm2, cultivated for 72 h and the single-cell-derived microclones were analysed as described below.
2.2. Time-lapse monitoring of movement of individual cells
The movement of A549 cells was time-lapse recorded for 12 h for the microclone approach or 4 h for the confluent monolayer approach, using a Leica DM IRE2 microscopic time-lapse system equipped with a ×20, NA (numerical aperture) 0.75 PlanApo objective, Leica DC350FX camera and Leica FW4000 software. The tracks of individual cells were determined from the series of changes in the cell centroid positions, as previously described (Daniel-Wojcik et al., 2008). Estimated cell movement parameters included: (i) the TLCD (total length of cell displacement; μm), i.e. the distance from the starting point directly to the cell's final position, (ii) VCD (velocity of cell displacement; μm/h), i.e. the distance from the starting point directly to the cell's final position/time of recording (4 or 12 h) and (iii) the VCM (velocity of cell movement; μm/h), i.e. total length of cell trajectory/time of recording (4 or 12 h). Cell trajectories from no less than 3 independent experiments (number of cells >50) were taken for estimation of statistical significance by the non-parametric Mann–Whitney test (#P<0.01).
2.3. Immunocytochemistry and immunoblotting
For immunocytochemical analyses of intracellular localization of Cx43 (connexin 43) and N-cadherin, cells were fixed with methanol/acetone (7:3, −20°C), labelled with rabbit anti-Cx43 and mouse anti-N-cadherin IgG (Sigma), stained with Alexa Fluor® 546-conjugated goat anti-rabbit and Alexa Fluor® 488-conjugated goat anti-mouse IgG (Invitrogen) and counterstained with 0.5 μg/ml bisbenzimide (Hoechst, Sigma; Czyz et al., 2005). Visualization of specimens was performed with a Leica DM IRE2 microscope equipped with ×40, NA 1.25 HCX PlanApo objective, Leica DC350FX camera and Leica FW4000 software.
3. Results and discussion
3.1. Heterogeneity of A549 cell populations
Microclone analyses showed the existence of 2 morphological subsets of A549 cells. ‘Epithelioid’ clones were characterized by strong intercellular adhesion resulting in a relatively low (<10%) fraction of physically isolated cells (Figure 1A) and low cell motility (Figure 1B). In contrast, ‘fibroblastoid’ clones consisted of physically isolated rear-front polarized (Figure 1D) and highly motile cells (Figure 1E). Over 70% of fibroblastoid clone cells (Figure 1F) but <10% of epithelioid cells (Figure 1C) displayed displacement rates higher than 4 μm/h. These data support previous observations on the heterogeneity of cancer cell lines in general (Langley and Fidler, 2007; Wysoczynski et al., 2007; Blick et al., 2008; Uchino et al., 2010) and on the co-existence of distinct cell sub-populations within the A549 cell line (Croce et al., 1999; Watanabe et al., 2002).
The question arises whether the identified cellular subsets belong to a single metastable population or represent 2 separate cell populations. Several observations support the first scenario. First, heterogeneity of clones with regard to the expression of N-cadherin could be observed (Figure 1G). Approximately 65% of epithelioid and 48% of fibroblastoid clones expressed this cell adhesion protein. Since N-cadherin is a marker of EMT (Haass et al., 2004), its heterogeneous expression in epithelioid and fibroblastoid cells may suggest an incomplete transition between 2 metastable states that results in the formation of ‘hybrid’ phenotypes. Similarly, both identified A549 cellular subsets expressed Cx43. Cx43 is a membrane protein constituting connexons and gap junctional channels that mediate gap junctional intercellular coupling, but are also involved in the regulation of cancer cell migration and invasion (Pollmann et al., 2005; Bates et al., 2007; Omori et al., 2007). However, plasmalemmally localized Cx43-positive plaques could be seen in the majority (>90%) of epithelioid clones (Figure 1G, arrow), but only within a relatively small fraction (31%) of fibroblastoid clones (Figure 1H). This observation indicates the heterogeneity of Cx43 functional status within A549 cell sub-populations. Secondly, fibroblastoid cells which comprised ∼60% of all analysed clones were characterized by higher mitotic activity than epithelioid cells, as concluded from differences in average numbers of cells in clones (12.5 and 9.9 for fibroblastoid and epithelioid clones respectively). The differences in proliferation rate between A549 cell subsets indicate that phenotypic transitions of fibroblastoid cells maintain A549 cell populations in a steady state. In the opposite case, relatively quiescent epithelioid cells would be eliminated from the culture due to their slower proliferation. Last but not least, a subclass of fibroblastoid clones characterized by the presence of cluster(s) of tightly attached cells (Figure 1H, arrow) could be discriminated. Such ‘intermediate’ clones were characterized by lower average cell number per clone than purely fibroblastoid clones (11.6 and 14.7 respectively compared with 9.9 estimated for epithelioid clones, Figure 1I). Their existence indicates occasional phenotypic switching of fibroblastoid cells giving rise to cells of the epithelioid phenotype. In general, the data suggest that proliferating fibroblastoid cells can revert to the quiescent epithelioid phenotype thus maintaining the A549 cell line and its heterogeneity.
3.2. Phenotypic characteristics of invasive A549 cells
To elucidate the invasive properties of A549 cell subsets and address directly the question of their metastability, we concentrated on the phenotype of the direct progeny of the cells that most readily transmigrated microporous membranes (A549T cells). Microclone analyses indicated considerable enrichment in cells of non-epithelioid phenotype, in particular fibroblastoid cells (Figure 2A). This is unsurprising since fibroblastoid cells displayed prominent motile activity, a prerequisite of transmigration. A correlation between the abundant fraction of fibroblastoid cells and increased abundance of motile cells in confluent monolayers of A549T1 populations was observed (Figure 2B). Furthermore, the elastic properties of fibroblastoid cells may predestine them to transmigrate through narrow obstacles (Friedl and Wolf, 2010). Interestingly, both fibroblastoid microclones (Figure 2C) and monolayer cultures of the A549T1 subset were enriched in cells displaying plasmalemmal localization of Cx43 (Figure 2E cf. Figure 2D for control A549 cells). This observation suggests that plasmalemmal Cx43 facilitates the transmigration of fibroblastoid A549 cells. A549 cells were seeded on the microporous membranes at a density securing their physical isolation. Therefore, plasmalemmal Cx43 could affect their transmigration potential through an effect on cancer cell adhesion and motility (Elias et al., 2007; Cronier et al., 2009), independently of gap junctional channel formation (Bates et al., 2007). However, such a pre-selection resulted in the formation of an A549 sub-population competent for gap junctional coupling (results not shown), which is a parameter crucial for penetration of natural barriers such as endothelium (Czyz, 2008).
3.3. Dynamics of phenotypic transitions within A549 cell populations
Further analyses of the heterogeneity of A549T populations demonstrated shifts in the relative numbers of fibroblastoid, intermediate and epithelioid cells within the propagated progeny of transmigrating A549 cells. Whereas an increase of the fibroblastoid fraction, and depletion of intermediate and epithelioid fraction, was observed at the initial stages of A549T cell propagation (passages 1 and 2, T1–T2), this tendency was reversed at later propagation stages (passages 3 and 4, T3–T4; Figure 2F). This observation indicates quiescence and/or senescence of epithelioid cells that accompanied the cells of fibroblastoid phenotype during micropore transmigration. Initially, their loss was not somehow compensated by phenotypic switches of fibroblastoid cells. The reappearance of the more abundant epithelioid sub-population (an 8–22% increase in their fraction) at later propagation stages may be due to a reduced loss of epithelioid cells and/or more intensive phenotypic transitions of fibroblastoid cells. Interestingly, the initially decreasing fraction of N-cadherin-positive epithelioid and intermediate clones (from 53 to 34% and 56.9 to 25.8% respectively) started to increase (to 61 and 42% respectively) at later stages of A549T cell propagation (Figure 2G). It can be concluded that both N-cadherin-positive and N-cadherin-negative fibroblastoid cells give rise to the epithelioid phenotype; however, a fraction of initially N-cadherin-negative intermediate and quiescent epithelioid cells switches on N-cadherin expression at later stages of their phenotypic transition. These observations directly confirm that fibroblastoid cells undergo an epigenetic programme resulting in the phenotypic reversal and the formation of heterogeneous epithelioid A549 cell populations.
In summary, we suggest that heterogeneous and invasive fibroblastoid A549 cells represent a ‘hybrid’ phenotype characterized by both epithelial and mesenchymal traits, and that they are capable of phenotypic switching. Their mitotic potential and capability of phenotypic transitions towards an epithelioid phenotype may indicate that they retain epigenetic memory and remain in 2 metastable states biased either towards self-renewal or ‘differentiation’ as already suggested for stem cells (Graf and Stadtfeld, 2008). These phenomena are crucial for cancer progression, and therefore A549 cell populations hold the promise of directing us towards key mechanisms of lung cancer metastatic dissemination.
Iga Bechyne took an active part in the experimental development of the study, and performed the experiments and statistical analyses. Katarzyna Szpak helped with cell cultures, discussion of results and paper drafting. Zbigniew Madeja participated in critical discussion of the results. Jarosław Czyż was responsible for the co-ordination, design and experimental development of the work and wrote the paper.
The Faculty of Biochemistry, Biophysics and Biotechnology of the Jagiellonian University is a beneficiary of structural funds from the
Berx, G, Raspe, E, Christofori, G, Thiery, JP and Sleeman, JP (2007) Pre-EMTing metastasis? Recapitulation of morphogenetic processes in cancer. Clin Exp Metastasis 24, 587-97
Blick, T, Widodo, E, Hugo, H, Waltham, M, Lenburg, ME and Neve, RM (2008) Epithelial mesenchymal transition traits in human breast cancer cell lines. Clin Exp Metastasis 25, 629-42
Croce, MV, Colussi, AG, Price, MR and Segal-Eiras, A (1999) Identification and characterization of different subpopulations in a human lung adenocarcinoma cell line (A549). Pathol Oncol Res 5, 197-204
Czyz, J, Guan, K, Zeng, Q and Wobus, AM (2005) Loss of beta1 integrin function results in upregulation of connexin expression in embryonic stem cell-derived cardiomyocytes. Int J Dev Biol 49, 33-41
Daniel-Wojcik, A, Misztal, K, Bechyne, I, Sroka, J, Miekus, K, Madeja, Z and Czyz, J (2008) Cell motility affects the intensity of gap junctional coupling in prostate carcinoma and melanoma cell populations. Int J Oncol 33, 309-15
Hugo, H, Ackland, ML, Blick, T, Lawrence, MG, Clements, JA and Williams, ED (2007) Epithelial–mesenchymal and mesenchymal–epithelial transitions in carcinoma progression. J Cell Physiol 213, 374-83
Omori, Y, Li, Q, Nishikawa, Y, Yoshioka, T, Yoshida, M, Nishimura, T and Enomoto, K (2007) Pathological significance of intracytoplasmic connexin proteins: implication in tumor progression. J Membr Biol 218, 73-7
Pollmann, MA, Shao, Q, Laird, DW and Sandig, M (2005) Connexin 43 mediated gap junctional communication enhances breast tumor cell diapedesis in culture. Breast Cancer Res 7, R522-34
Sottoriva, A, Verhoeff, JJ, Borovski, T, McWeeney, SK, Naumov, L and Medema, JP (2010b) Cancer stem cell tumor model reveals invasive morphology and increased phenotypical heterogeneity. Cancer Res 70, 46-56
Uchino, M, Kojima, H, Wada, K, Imada, M, Onoda, F and Satofuka, H (2010) Nuclear beta-catenin and CD44 upregulation characterize invasive cell populations in non-aggressive MCF-7 breast cancer cells. BMC Cancer 10, 414
Watanabe, N, Dickinson, DA, Krzywanski, DM, Iles, KE, Zhang, H and Venglarik, CJ (2002) A549 subclones demonstrate heterogeneity in toxicological sensitivity and antioxidant profile. Am J Physiol Lung Cell Mol Physiol 283, L726-36
Wysoczynski, M, Miekus, K, Jankowski, K, Wanzeck, J, Bertolone, S and Janowska-Wieczorek, A (2007) Leukemia inhibitory factor: a newly identified metastatic factor in rhabdomyosarcomas. Cancer Res 67, 2131-40
Received 18 March 2011/20 July 2011; accepted 13 September 2011
Published as Cell Biology International Immediate Publication 13 September 2011, doi:10.1042/CBI20110151
© The Author(s) Journal compilation © 2012 Portland Press Limited
ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)