The human RB cell line RB-355, established and first described by Griegel et al (1990) (8), and formerly donated by K. Heise, was kindly provided by Dr H. Stephan. The RB cell lines Y-79 (9) and WERI-Rb1 (10), originally purchased from the Leibniz Institute DSMZ (German Collection of Microorganisms and Cell Cultures), were likewise kindly provided by Dr H. Stephan. All RB cell lines were last tested and authenticated in September 2015. Mutation analyses were conducted using an MLPA kit (SALSA MLPA kit P047 RB1; MRC-Holland, Amsterdam, The Netherlands) and reactions were performed according to the manufacturer's instructions. Additional sequencing of the RB1 gene was performed for all RB cell lines. However, most recent STR analyses (March 2017) confirmed the authenticity of the cell lines.

The cell lines were cultivated as suspension cultures in Dulbecco's modified Eagle's medium (DMEM) with 15% fetal bovine serum (FBS) (both from PAN-Biotech GmbH, Aidenbach, Germany), 100 U penicillin/ml and 100 µg streptomycin/ml, 4 mM L-glutamine (both from Gibco, Karlsruhe, Germany), 50 µM β-mercaptoethanol (Carl Roth, Karlsruhe, Germany) and 10 µg insulin/ml (PAN-Biotech) at 37°C, 10% CO₂ and 95% humidity. No approval from an Ethics Committee was required for work with the human cell lines.

All chemoresistant RB cell lines characterized were generously provided by Dr H. Stephan. To generate these cell lines, established Y-79, WERI-Rb1 and RB-355 cells (see above) were continuously treated with consecutively increasing concentrations of etoposide or cisplatin (both from Teva, Berlin, Germany) until the chemoresistant sublines exhibited a at least 10-fold higher IC₅₀ value in WST-1 viability assays than the respective parental controls (11). The chemoresistant cell lines were subsequently cultivated as described above for RB cell lines with additional treatment of the appropriate cytostatic drug twice a week (every 3–4 days). For details on final concentrations of the drugs used, see Table I.

Cell proliferation was determined by 5-bromo-2′-deoxyuridine (BrdU; Sigma, Steinheim, Germany) incorporation. For BrdU immunocytochemistry 10 µM BrdU was added to the cells 4 h prior to paraformaldehyde (PFA; Sigma) fixation. Cells were incubated with a rat anti-BrdU antibody (1:1,000; ab6326; Abcam, Cambridge, UK) and proliferating cells were visualized using a goat anti-rat antibody labelled with Alexa Flour^® 488 (1:1,000; A11006; Life Technologies, Darmstadt, Germany). In order to determine changes in apoptosis levels, cells were stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma) or Click-iT^® Plus TUNEL assay for in situ apoptosis detection (cat. #C10617; Thermo Fisher Scientific, Darmstadt, Germany), following the manufacturer's protocol and pycnotic nuclei were manually counted as previously described by our group (12).

To determine growth kinetics, 3×10⁵ RB cells were seeded in 500 µl DMEM with supplements in a 24-well plate and vital cells were counted manually using the trypan blue exclusion method. Cells were seeded in triplicates and counted at several time points (24, 48, 72 and 96 h). In the case of the extremely slow-growing Y-79 cisplatin-resistant RB cell line, we recorded long-term growth curves over a period of 336 h with longer counting intervals and plotted the values logarithmically to visualize differences between resistant and control cells.

Soft agarose assays were performed as previously described in detail (13). The colony formation efficiency (%)/visual field was determined by counting the colonies and single cells in 5 visual fields (magnification, ×10)/cell line in triplicates and images were captured using a Leica DMIL or a Nikon Eclipse TS2 microscope equipped with a digital camera and ProgRes Capture Basic 1.2.01 or IC Capture 2.4 (The Imaging Source) software.

In order to test for changes in the migration and tumor formation capacity following etoposide and cisplatin resistance, RB cells were grafted onto the chick chorioallantoic membrane (CAM) as described in a recent publication by our group (14). Mainly following the metastasis model protocol published by Zijlstra et al (2002) (15), and visualized by Palmer et al (2011) (16), 50 µl cell suspension [1×10⁶ chemoresistant or control cells in phosphate-buffered saline (PBS)] was grafted onto the CAM area. For each RB cell line, 30 eggs were grafted (15 with parental and 15 with chemoresistant RB cells) in at least 3 independent experiments. All in vivo chick CAM experiments were conducted according to the relevant national guidelines of the responsible authority, the State Office for Nature, Environment and Consumer Protection (LANUV) as well as to the Directive 2010/63/EU of the European Parliament and of the council of September 22, 2010 on the protection of animals used for scientific purposes, which does not comprise any restrictions for the use of non-mammalian embryos. However, the Institutional Animal Care and Use Committee (IACUC) of the Medical Faculty of the University Hospital Essen approved the CAM assays and no ethical approval was required as according to the German Animal Experiment and Welfare Guidelines, ethical approval is only essential when animals are intended to live beyond hatching.

The duration of the chick CAM assay is limited to a 7–9 day window prior to hatching. Seven days after grafting (E10-17) chick embryos were anesthetized by cooling on ice and sacrificed by decapitation. CAM tumors were excised, measured, photographed and fixed for 1 h at 4°C in 4% PFA in 0.1 M phosphate buffer (pH 7.4). For cryo-embedding, the tumor tissue was incubated for 30 min in PBS (pH 7.3) containing 15% sucrose, followed by a 30-min incubation in PBS containing 30% sucrose and finally embedded in OCT compound (Tissue-Tek; Germany), and sectioned at 10 µm using a cryostat. Images and measurements of tumors forming on the upper CAM were captured with a Nikon stereo dissecting microscope SMZ 1000 equipped with a Nikon digital camera and Nikon EclipseNet software. Exemplarily, images were captured with a 3D-Profilomter VR-3200 microscope (Keyence, Neu-Isenburg, Germany) to visualize changes in the 3D volume of the tumors.

The localization of the tumors forming in the upper CAM after grafting was visualized on hematoxylin and eosin (H&E)-stained cryosections. The human origin of the tumors forming in the chicken CAM after grafting human RB cells was verified using a mouse anti-human nuclear antibody (MAB 128; Merck Millipore, Darmstadt, Germany) at a dilution of 1:100 in PBS containing 0.1% Triton, 4% bovine serum albumin (BSA) and 1% normal goat serum (NGS) overnight at 4°C. The reaction was visualized using a goat anti-mouse antibody labelled with Alexa Flour^® 488 (A11001; Molecular Probes, Camarillo, CA, USA), diluted 1:1,000 in PBS with 1% BSA for 2 h at room temperature.

For β-tubulin stainings, 1×10⁵ cells were stained on coverslips as previously described (17). In brief, cells were fixed with 4% PFA in PBS for 1 h following 3 washes with PBS. Afterwards, cells were incubated with ice-cold 100% methanol for 5 min on ice, washed 3 times with PBS and incubated for 1 h in blocking solution [PBS, 0.3% Triton, 4% BSA, 5% NGS (Dako, Hamburg, Germany)]. Anti β-tubulin primary antibody (T-4026; Sigma-Aldrich, Munich, Germany) was diluted 1:200 in PBS containing 0.1% Triton, 4% BSA and 1% NGS and cells were incubated overnight at 4°C. Following 3 washes with PBS, goat anti-mouse secondary Alexa 488-coupled antibody (Molecular Probes, Germany) was used at 1:1,000 dilutions in PBS/1% BSA (Roth). Finally, cells were counterstained with DAPI to visualize the nucleus. As controls, in all cases PBS was substituted for the primary antisera in order to test for non-specific labeling. No specific cellular staining was observed when the primary antiserum was omitted.

For measurements of cell and nuclear sizes, images from β-tubulin-stained cells on coverslips were acquired using a Nikon Eclipse E600 microscope equipped with a digital camera. For each cell line, 5 equally distributed, x-shape rendered visual fields/coverslip were captured, and the cytoplasmic and nuclear outline of 7 cells/visual field were measured using the Nikon Eclipse net measurement software.

All assays were performed at least in triplicate. Statistical analyses were conducted using GraphPad Prism 6. Data represent means ± SEM of 2 to 5 independent experiments from independent RB cell cultures. Results were analyzed by a Students t-test or one-way ANOVA and Newman-Keuls post hoc test and considered significantly different at *P<0.05, **P<0.01 or ***P<0.001. Statistics on the growth curves was performed using a free web interface http://bioinf.wehi.edu.au/software/compareCurves/, which uses the compareGrowthCurves-function from a statistical modeling package called ‘statmod’, available from the R Project for Statistical Computing: http://wwww.r-project.org, previously described elsewhere (18).

In order to check for morphological changes following etoposide and cisplatin resistance, we compared the appearance of parental and chemoresistant Y-79, WERI-Rb1 and RB-355 cells a) in cell culture (Fig. 1A) and b) after seeding on poly-D-lysine-coated coverslips and immunocytochemical staining with β-tubulin and DAPI counterstaining (Fig. 1B). While all parental controls and particularly WERI-Rb1 cells frequently form chain-like structures when seeded on coverslips, all chemoresistant RB cell lines tended to form clusters, which were more pronounced in etoposide-resistant cell lines than in cisplatin-resistant cells. However, the etoposide- and cisplatin-resistant cell lines seemed to become partially adherent and sometimes neurite-like processes could be observed.

Measurements of cell and nuclear size revealed that compared to their parental counterparts, cisplatin-resistant cells were significantly bigger and had larger nuclei (Fig. 2A and B), whereas etoposide-resistant cells did no exhibit significant changes in either size (data not shown).

The etoposide-resistant cell lines Y-79 and WERI-Rb1, established from two well-established RB cell lines, originally derived from unilateral RB tumors (10), exhibited significantly higher growth rates compared to the parental etoposide-sensitive control cells (Fig. 3A and B). In semi-adherent RB-355 RB cells, likewise derived from an unilateral RB tumor, but exhibiting different growth kinetics (19), cell growth was not significantly affected by etoposide resistance (Fig. 3C).

As revealed by WST-1 assays, etoposide resistance resulted in a significant increase in Y-79 cell viability. WERI-Rb1 etoposide-resistant cells displayed a slightly, but not significantly increased viability, whereas etoposide resistance did not significantly alter the viability of RB-355 cells compared to the parental controls (Fig. 3D).

Cell proliferation was significantly increased in all three etoposide-resistant RB cell lines analyzed as reflected by significantly higher numbers of BrdU-positive cells (Fig. 3E).

Etoposide treatment still induces apoptosis in etoposide-resistant RB cell lines. As revealed by DAPI cell counts (Fig. 3F) and confirmed by TUNEL assays (Fig. 4A) continuous treatment with etoposide still significantly induced apoptosis in the WERI-RB1 etoposide-resistant RB cells, whereby the increase in cell death levels in the Y-79 and RB-355 etoposide-resistant cell lines was not significant. The remaining pro-apoptotic effect of etoposide seemed to counterbalance its pro-proliferative effect after induction of resistance (Fig. 3E) resulting in an absent overall effect of etoposide on cell viability in the WERI-Rb1 and RB-355 etoposide-resistant cells (Fig. 3D). By contrast, in Y-79 etoposide-resistant cells, displaying significantly increased cell viabilities (Fig. 3D), the strong pro-proliferative effect of etoposide-resistance appeared to predominate (Fig. 3E).

Photo-documentation (Fig. 5A and B), counts (Fig. 5C) and measurements of the tumors (Fig. 5D and E) developing from Y-79, WERI-Rb1 and RBL-355 cells grafted on the upper CAM revealed that the tumor formation capacity of Y-79 and WERI-Rb1 etoposide-resistant cell lines was significantly increased compared to their parental counterparts (Fig. 5C). Tumor formation capacity of semi-adherent RB-355 etoposide-resistant cells likewise increased, but did not reach significance. However, the etoposide-resistant RB cell lines developed larger tumors (Fig. 5D) and tumors of significantly higher weight (Fig. 5E) when compared with the chemosensitive control cells.

The localization of the tumors developing in the upper chicken CAM at the border between CAM ectoderm and mesoderm 7 days after grafting human RB cells was visualized by H&E staining of tumor cryosections (Fig. 6A). The human nature of the tumors was verified immunocytochemically, using an anti-human nuclear antibody (Fig. 6B-D).

Etoposide resistance changed the anchorage independent growth of Y-79 and WERI-Rb1 cells as reflected by a significant decrease in their colony formation capacity in soft agarose (Fig. 7A). However, all three etoposide-resistant cell lines analyzed formed considerably smaller colonies when compared with the chemosensitive control cells (Fig. 7B).

All cisplatin-resistant cell lines analyzed displayed significantly lower growth kinetics compared to the control cells (Fig. 8A-C). Among the three cisplatin-resistant RB cell lines investigated, Y-79 cells were the slowest in growth and thus, we had to record long-term growth curves over a period of 336 h (instead of 96 h) and plot the values logarithmically to visualize differences between resistant and control cells (Fig. 8A). As revealed by WST-1 assays cisplatin resistance likewise resulted in a significant reduction in cell viability of all three cisplatin-resistant cell lines (Fig. 8D). Cell proliferation was significantly decreased in the Y-79 and WERI-Rb1 cisplatin-resistant RB cell lines as reflected by significantly lower numbers of BrdU-positive cells (Fig. 8E), but remained unchanged in the RB-355 cisplatin-resistant cells (Fig. 8E).

As revealed by DAPI cell counts (Fig. 8F) and confirmed by TUNEL assays (Fig. 4B) cisplatin treatment still significantly increased the apoptosis levels in the cisplatin-resistant RB cell lines.

Photo-documentation (Fig. 9A and B) and measurements revealed that tumors developing from Y-79 and WERI-Rb1 cisplatin-resistant RB cells were significantly smaller (Fig. 9C). Y-79 cisplatin-resistant cells likewise exhibited significantly lower weights than tumors developing from chemosensitive control cells (Fig. 9D), whereby the same tendency did not reach significance in the WERI-Rb1 cisplatin-resistant cells. Compared to the respective parental controls, tumors forming from RB-355 cisplatin-resistant cells neither decreased nor increased in size and weight (Fig. 9C and D). Cisplatin resistance did not significantly influence the tumor formation capacity of the three cell lines investigated (data not shown).

Soft agarose assays revealed that cisplatin resistance significant decreased the colony formation capacity of all cisplatin-resistant RB cell lines analyzed (Fig. 10A). However, all cisplatin-resistant cell lines formed considerably smaller colonies compared to the control cells (Fig. 10B).