3. Discussion
The RPE forms the outer blood-retinal barrier and plays an essential role in visual function [18]. Since it is located between the choroids and the neurosensory retina, it has to fulfill important functions, like absorption of light, reisomerization of all-trans retinal into 11-cis retinal, protection against photooxidation, epithelial transport of ions, nutrients and fluids, phagocytosis of photoreceptor outer segments, secretion of essential factors for the integrity of neighboring tissues and supporting the immune privilege of the inner eye [19]. As one characteristic of ERU is the breakdown of the outer blood-retinal barrier, our aim is to elucidate the pathomechanisms that are involved in this breakdown. Therefore, performing functional studies on equine RPE cells will be necessary to understand their role in ERU [20]. We set our focus on cell surface membrane proteins in this study, as we expect them to be targets for initiating events in the breakdown of the outer blood-retinal barrier in eyes of ERU diseased horses. In order to detect differentially regulated cell surface membrane proteins in the eyes of affected horses, it is a prerequisite to identify the cell surface membrane proteome of equine RPE cells in the physiological state. We compared native equine RPE cells with passage-4 cells to determine if there were any changes in this proteome due to cultivation and dedifferentiation. Changes in the whole proteome without specific focus on cell surface membrane proteins in cultured human RPE cells compared to native human RPE cells were described previously [14]. To our knowledge, we are the first to describe the cell surface membrane proteome of native and cultured equine RPE.
A total of 112 proteins were identified by LC MS/MS, of which 84% were cell membrane proteins (Table 1). Twenty-three proteins were concurrently expressed in both, native and passage-4 RPE cells which equals 20.5% of all 112 identified proteins (Table 1, proteins 1–23). Whereas 28 proteins (25% of all 112 identified proteins) were exclusively expressed in native cells (Table 1, proteins 24–51), 61 proteins (54.5% of all 112 identified proteins) were expressed only in passage-4 cells (Table 1, proteins 52–112), indicating a profound re-organization of the cell surface proteome in cultured equine RPE cells.
As expected, beta-actin, an ubiquitously expressed housekeeping protein [21], was equally expressed in both examined states (Table 1, protein 3). Among the equally expressed proteins, six cytokeratins were present, which are intermediate filament proteins (Table 1, proteins 7–12). Desmoplakin, a desmosomal protein which mediates intercellular adhesion in human RPE cells [22], was also equally expressed in both states of equine RPE cells (Table 1, protein 13), as well as basigin (Table 1, protein 2), which plays a key role in extracellular matrix (ECM), remodeling [23] and neuroplastin (Table 1, protein 18), which belongs to the same protein family [24]. Basigin was already described as being expressed by human RPE in situ and by a human RPE cell line, namely ARPE-19 [25]. Further, four different types of Na+/K+ ATPases were identified, which are usually localized apically in RPE cells [26]. Three of them were expressed in both states (Table 1, proteins 15–17), whereas one of them was only present in passage-4 RPE cells (Table 1, protein 98).
Among the 28 proteins exclusively expressed in native equine RPE cells, there were three cytokeratins (Table 1, proteins 31–33) and few ion-channel proteins, namely Kir7.1; solute carrier family 1, member 4; solute carrier family 12, member 2; solute carrier family 13, member 3; solute carrier family 16, member 1; solute carrier family 4, member 7; solute carrier family 6, members 6, 9, 13, 20; and solute carrier organic anion transporter family, member 1A2 and member 1B3 (Table 1, proteins 39–49). Further, we identified two RPE cell markers, CRALBP and RPE65 [16] (Table 1, proteins 29 and 38). CRALBP, RPE65 and also RDH5, which was exclusively expressed by native RPE cells (Table 1, protein 37), are proteins playing a key role in the visual cycle [27–29]. CRALBP, RPE65 and RDH5 were abundant in native and absent in cultured RPE cells. Correlating with our results, former research observed that RPE65 and CRALBP, both proteins associated with highly specialized functions of the RPE, were absent in cultured RPE cells [14,30]. Regarding that CRALBP is not only expressed by RPE cells, but also, for instance, by Mueller glial cells [31], we believe that RPE65 is an excellent marker for the verification of native in comparison to cultured RPE, as it was found to be differentially expressed by Western blot analysis, immunocytochemistry and even PCR (Figure 2–4). Another reason for considering RPE65 an attractive protein for further research in ERU is the fact that immunoreactions against RPE65, next to CRALBP and S-Antigen, are known to induce experimental autoimmune uveitis in rats [7,32,33].
Besides cell surface proteins, which were only expressed by native equine RPE cells (Table 1, proteins 23–51), there were 61 proteins exclusively detected in cultured passage-4 RPE cells (Table 1, proteins 52–112). Interestingly, there was a notable increased expression of several clusters of differentiation (CD) molecules in passage-4 cells (Table 1, proteins 59–79) compared to the native state, whereas ion channel protein expression decreased (Table 1). An altered expression of ECM proteins could be observed as well, for instance of fibulin 1, fibulin 2, fibronectin and thrombospondin 1. (Table 1, proteins 85–87 and 108). Thrombospondin 1 is not only an ECM protein expressed by RPE cells, but it is also known to suppress activated T-cells [34], which additionally makes this protein very interesting for ERU research. It was found to be exclusively expressed in cultured equine RPE cells. Fibronectin was also absent in native equine RPE cells, but abundant in passage-4 cells (Table 1, protein 85; Figures 3 and 4). Immunocytochemically, fibronectin shows positive immunoreactivity with a perinuclear expression pattern in passage-4 cells, and no immunoreactivity was observed in native RPE cells (Figure 4). In vivo, the RPE basement membrane contains ECM proteins such as laminin, fibronectin, vitronectin and collagen type VI [35]. When RPE is dissociated from the eyecup, the RPE basement membrane remains in the eyecup, and there is no matrix for the RPE cells to attach to in vitro. In order to adhere and form a monolayer of epithelial cells in vitro, RPE cells begin to produce ECM proteins, such as fibronectin, themselves [36,37]. The perinuclear expression pattern in passage-4 indicates an augmented protein synthesis of the ECM protein fibronectin in passage-4 cells. Furthermore, there was an altered expression of proteins associated with cell adhesion, predominantly integrins (Table 1, proteins 69–72, 76 and 92–95), that are known to not only interact with ECM proteins [38], but also cadherins, catenin, cell adhesion molecule 1, dystroglycan 1, vitronectin and versican (Table 1, proteins 55, 56, 58, 73, 80, 81 and 112). CD49c, which is also known as integrin-alpha 3, belongs to a family of receptors for ECM and cell adhesion molecules [39]. Gullapalli et al. observed an increased expression of CD49c on transcriptomic and proteomic level in cultured human RPE cells [40], whereas uncultured human RPE cells did not express or only expressed low amounts of CD49c. These findings were consistent with their real-time PCR results [40]. We were able to demonstrate an altered expression in passage-4 cells as well, compared to native cells on mRNA and proteomic level (Table 1, protein 70; Figures 2–4). In Western blot analysis, no signal in the lane with native RPE cell lysate was detectable, whereas the lane with passage-4 cell lysate showed a distinct signal. Also, no staining for CD49c could be shown in native cells by immunocytochemistry, confirming CD49c expression on proteomic level to be specific for passaged RPE cells (Figures 2–4).
Interestingly, we showed that the cell surface proteome of equine RPE cells shifted from proteins with RPE specialized functions (CRALBP, RPE65 and RDH5) in native cells (Table 1, proteins 29, 37 and 38; Figures 2–4) to proteins that are associated with cell adhesion and ECM formation (integrins, cell adhesion molecule 1, catenin, thrombospondin 1, fibronectin) in cultured cells (Table 1, proteins 58, 80, 85, 108 and 93–95; Figures 2–4).
The altered expression of molecules that are associated with cell adhesion in passage-4 cells could be clearly demonstrated in Figure 1 (dark green segments). According to Figure 1A, in native equine RPE cells, only 2.29% of all respective proteins are associated with cell adhesion (dark green segment), whereas in passage-4 cells, 12.06% of the expressed proteins are capable of cell adhesion as their biological process (Figure 1B, dark green segment). These observations in equine cells correspond to the findings in cultured human cells where a shift to processes such as cell adhesion could be accordingly displayed [14]. Another notable increase in passage-4 RPE cells could be seen in cell communication (Figure 1A,B, grey segments). As the normal function of the RPE depends on the maintenance of tight adhesions between the cells [41], and cell-cell adhesion and intercellular communication are integrated [42], it is not surprising that this biological process alters concurrently in passage-4 RPE cells (Figure 1A,B, grey segments). Interestingly the relation between cell adhesion and communication remains equal in both states (Figure 1A,B, grey segments).
In contrast to alteration of proteins associated with cell adhesion and communication in passage-4 cells, Figure 1 demonstrates a noticeable decrease of proteins that are associated with transport as their biological process in passage-4 cells (6.74%) (Figure 1B, bright blue segment) compared to the native state (34.86%) (Figure 1A, bright blue segment), correlating with the observations in Table 1, where in passage-4 cells a strong decrease in ion transporting channels was noted.
Our results therefore clearly demonstrate that there is a profound re-organization of the cell surface membrane proteome in cultured equine RPE cells compared to native RPE cells. We believe that one reason for these differences is most probably the in vitro situation, which profoundly differs from the situation in vivo, where RPE is provided with environmental specific factors by surrounding tissue that are not included in standard cell culture media [43]. The lack of the choroid on the basal side and the retina on the apical side of the cells in culture could lead to a decrease of functions like transport, which is not needed in its entirety in vitro. Then again, other functions, as cell-adhesion, become more important. Since retinal tissue is destroyed and retinal architecture is widely disintegrated in ERU [13], we assume that the proteomic changes in diseased eyes are very similar to the changes observed in this study, as disintegrated and destroyed retinae could be compared to a missing retina in culture. In some ways, culture is like a disease state, where a degenerate neural retina no longer sends RPE the signals that modulate key functions [11]. Another reason for the differences in cell surface proteomes of native and cultured RPE cells is that RPE cells do not divide in vivo, in contrast to RPE cells in vitro[44]. All of this may lead to an altered expression of proteins, as the proteome is dynamic and changes with the physiological state of the cell [14]. In the past, changes in protein expression due to in vitro culturing and passaging, could be observed in cells other than RPE cells as well [45,46]. Our results clearly validated this and demonstrated that, by culturing equine RPE cells, changes in the cell surface proteome occur and that the passage-4 cells only to a limited extend show properties of native physiological RPE cells.
Regarding the short time post mortem until further processing of the equine eyes used in this study in comparison to eyes obtained from human donors, we were able to perform analyses on native RPE cells without loss of proteins due to degradation. By means of biotinylation and LC MS/MS in order to capture, enrich and identify cell surface membrane proteins, we were the first to characterize the cell surface proteome of RPE cells of two different states, native and passage-4. Detecting a total of 28 CD molecules, we contributed to a profile and repertoire of RPE CD antigens of which, to our knowledge, ten have not yet been described to be expressed in RPE cells of any species before (Table 1, proteins 25–27, 59, 60 and 64–67). We believe that a cell surface proteome of healthy equine RPE cells is essential and a necessity in order to compare RPE cells of healthy physiological and diseased pathological state in the future, and thus provides the basis for further research in this field. Furthermore, knowledge of the cell surface proteome of equine RPE cells is not only of great value for research of ERU, but also of major importance for research in the human field of other severe eye diseases with participation of the RPE as ocular tissue, namely Age Related Macular Degeneration (AMD), human autoimmune uveitis and Proliferative Vitreoretinopathy (PVR) [47–49]. Since ERU is the only spontaneous model for human autoimmune uveitis [50], it is of great importance to investigate and understand the pathomechanisms leading to this severe disease. The lack of proteins with highly specialized RPE functions in cultured RPE compared to native cells requires the use of native cells in research. Therefore, we believe that primary cells should be used for experiments in which it is important to evaluate changes in these specialized proteins in healthy versus diseased RPE cells.
Since cultured cells show a great tendency to form intercellular adhesions, these cells could be used for functional studies where tight junctional formation is required, for example as transmigration assays on semi-permeable filter systems. In regard to availability, cultured cells have a significant advantage for the use in research compared to native cells. Further, they can be passaged, and thus multiplied, enabling functional studies to be performed with a much higher amount of cells and a higher diversity of experimental setups.
Nevertheless, future proteomic studies on diseased equine RPE cells have to be conducted to clarify if the shift of the cell surface proteome in cultured equine RPE cells to proteins, which were not detectable in native cells, could be similar to equine RPE cells in diseased state, as was already shown for protein expression profile of dedifferentiated human RPE cells and initial stages of PVR [14].
To summarize, we were the first to provide a characterization of the cell surface proteome of native and cultured equine RPE cells by means of cell surface protein biotinylation and high resolution mass spectrometry. Additionally, we showed that there is a considerable difference between native and cultured equine RPE cells regarding their cell surface protein expression by investigation of differentially expressed proteins. Additionally, we were the first to describe ten CD molecules in equine RPE cells, which, to our knowledge, have not yet been described in the context of RPE cells of any species (Table 1, proteins 25–27, 59, 60 and 64–67). Cultured RPE cells experience a profound re-organization of the cell surface proteome since only 20.5% of all 112 identified proteins were expressed equally in both cell states. An interesting question that arises is how quickly the proteome of equine RPE cells changes after preparation of native cells and subsequent cultivating. As we set our focus on the comparison between native RPE cells and cells of passage-4, future experiments need to be conducted to evaluate if and when this proteomic shift or a tendency is notable in passage-2 and -3 cells. Nevertheless we believe that we have set a great foundation with regard to future experimental studies on the function of RPE cells in physiological and pathological conditions.
4. Experimental Section
4.1. Preparation of Equine Retinal Pigment Epithelium Cells
No experimental animals were used in this study. Eyes providing healthy equine retinal pigment epithelium samples were obtained at a local abattoir. Eyes were considered normal based on the diagnosis of the veterinarian present at the abattoir, medical histories of the horses as provided by their owners and preliminary histological analysis. The collection and use of equine eyes from animals that were killed due to a research-unrelated cause, was approved for purposes of scientific research by the appropriate board of the veterinary inspection office Munich, Germany (Permit number: 8.175.10024.1319.3). RPE cells from eyes of eight different eye-healthy horses were used in this study, four biological replicates of native cells and four biological replicates of passage-4 cells. For each verification procedure, we had at least two technical replicates (two for PCR and immunocytology, four in Western blots).
RPE cells were prepared as follows: First, residual periocular tissue was carefully removed with forceps and scissors. Then, the intact eyes were rinsed in 70% Ethanol for 2 min followed by a short washing step in cold phosphate buffered saline (PBS). Afterwards, eyes were stored in sterile Dulbecco’s Modified Eagle Medium (DMEM) until further processing. All following preparation steps were performed under a laminar flow hood with sterile instruments. Eye globes were cut open circumferentially, and anterior parts of the eye, vitreous and neurosensory retina, were carefully removed. Posterior eyecups were rinsed twice with 37 °C pre-warmed PBS to wash clear of residual vitreous and retinal tissue without destroying the RPE layer. For complete removal of retinal tissue, especially the rod outer segments, eyecups were filled with pre-warmed 1mM PBS/EDTA pH 7.4 and incubated at room temperature (RT) for 15 min. The PBS/EDTA solution was discarded and, in order to enzymatically detach the RPE cell layer, eyecups were filled with pre-warmed dissociation buffer containing 1 μL papain (CellSystems, Troisdorf, Germany) per milliliter buffer (1 mM PBS/EDTA pH 7.4, 260 mM l-Cysteine, Sigma Aldrich, Deisenhofen, Germany, 1% BSA, AppliChem, Darmstadt, Germany). The eyecups were incubated at 37 °C and 5% CO2 atmosphere for 25 min. Then, the RPE cells were resuspended in dissociation buffer within the eyecup. The enzymatic activity of papain was blocked by transferring the suspension containing the RPE cells into DMEM supplemented with 10% heat-inactivated fetal calf serum (FCS) and 1% penicilline/streptomycine (P/S). The cell suspension was centrifuged at RT, 130× g for 5 min, and the resulting cell pellet was carefully washed twice with pre-warmed PBS. Finally, the RPE cells were either seeded into T25 cell culture flasks (Sarstedt, Nümbrecht, Germany), or biotinylated and processed for mass spectrometry. For passage-4 cells, RPE cells were cultured in DMEM supplemented with 10% heat-inactivated FCS and 1% P/S. The cells were maintained at 37 °C and 5% CO2 atmosphere until they reached confluence after 14 days of cultivation. For protein expression analysis by mass spectrometry, cells were harvested with Trypsin/EDTA (Biochrom, Berlin, Germany), washed twice with cold PBS and centrifuged in between washing steps at 4 °C, 500× g for 10 min. Prior to cell lysis, cell surface membrane proteins of 1.6 × 106 passage-4 RPE cells and 5.0 × 105 native RPE cells were labeled with 542 μg biotin and 50 μg biotin, respectively. After 30 min rotation at 4 °C and two centrifugation steps at 2800× g and 16,000× g for 10 min each at 4 °C, cells were lysed with 1% Nonidet P-40, 150 mM NaCl, 1 × Roche Complete Protease Inhibitor, EDTA-free; 5 mM 2-Iodacetamide in 10 mM Tris-HCl pH 7.6 and biotinylated cell surface proteins were captured using Streptavidin-beads (IBA, Goettingen, Germany). After extensive washing to remove non-specifically bound proteins, captured proteins were cleaved through digesting beads overnight with trypsin (Promega, Mannheim, Germany) followed by incubation with glycerol-free PNGase F (New England Biolabs, Frankfurt/Main, Germany) at 37 °C. Finally, peptides were analyzed by LC-MS/MS.
4.2. Mass Spectrometry
LC-MS/MS mass spectrometry was performed as previously described [1,5]. Briefly, cell lysates were subjected to on-membrane trypsin digest. Resulting peptides were separated on a reversed phase chromatography column (PepMap, 15 cm × 75 μm ID, 3 μm/100A pore size, LC Packings), which was operated on a nano-HPLC apparatus (Ultimate 3000, Dionex, Idstein, Germany) connected to a linear quadrupole ion-trap Orbitrap (LTQ Orbitrap XL, Thermo Fisher Scientific, Schwerte, Germany). The mass spectrometer was operated in the data-dependent mode to automatically switch between Orbitrap-MS and LTQ-MS/MS acquisition. Survey full scan MS spectra (from m/z 300 to 1500) were acquired in the Orbitrap resolution R = 60,000 at m/z 400. The method used allowed parallel selection of up to the ten most intense ions for fragmentation on the linear ion trap using collision induced dissociation at a target value of 100,000 ions and subsequent dynamic exclusion for 30 s. MS/MS spectra were exported from the Progenesis software as Mascot Generic file (mgf) and used for peptide identification with Mascot (version 2.3.02, Matrix Science: London, UK; available online: http://www.matrixscience.com, accessed on 26 June 2012) in the Ensembl database for horse (Equus caballus; EquCab2.56.pep, available online: ftp://ftp.ensembl.org/pub/current_fasta/equus_caballus/pep/, accessed on 26 June 2012) containing a total of 22641 protein sequences. A protein was considered as identified if the confidence score was higher than 30 at a significance threshold for the Mascot result of p ≤ 0.01.
4.3. Gene Ontology (GO)
Gene Ontology annotations were retrieved from Ensembl and EMBL-EBI QuickGO database. GO annotations were available for all identified proteins and were classified based on three organizing principles of GO, namely: biological process, cellular component and molecular function. GO annotations for each protein were hierarchically summarized using AmiGO. Based on these summarized GO annotations pie charts were made, illustrating differences in biological processes, cellular component and molecular function between native and passage-4 equine RPE cells. In this manuscript pie charts representing differences in biological processes of native and passage-4 RPE cells are presented. Biological processes, which are represented by 1.1% or less of all proteins, are summarized in the field “others” (Figure 1A,B).
4.4. Polymerase Chain Reaction (PCR)
RNA was isolated from freshly isolated equine RPE and cultured RPE of passage-4 using RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Schwerte, Germany), 2 μg of RNA were reverse transcribed in presence and absence of reverse transcriptase, according to the manufacturer’s protocol. PCR was then performed in a total volume of 50 μL using the cDNA as a template in the presence of specific primers for the respective proteins. Reaction buffer, dNTPs and Taq DNA Polymerase were obtained from Fermentas (St. Leon-Rot, Germany). The thermocycler, which was used for PCR, was a Tetrad PTC-225 Thermal Cycler (MJ research, St. Bruno (Quebec), Canada). PCR, which was performed on samples that were reverse transcribed to cDNA in the absence of reverse transcriptase, did not show any amplified product. Amplified products were separated by 2% agarose gel electrophoresis, stained with Serva DNA Stain G (Serva Electrophoresis GmbH, Heidelberg, Germany) and visualized with UV illumination (Herolab UVT-40M, Wiesloch, Germany) (Figure 2).
4.5. Western Blot Analysis
Retinal pigment epithelial cell pellets were solubilized in lysis buffer (9 M urea, 2 M thiourea, 1% DTT, 4% CHAPS, and 2.5 mM each of EGTA and EDTA), and protein content was quantified with Bradford assay (Sigma-Aldrich, Deisenhofen, Germany). Equal total protein amounts (5 μg) of RPE samples were resolved in 1D SDS gels and then blotted semidry onto polyvinyl-difluoride membranes (GE Healthcare, Freiburg, Germany). Unspecific binding was blocked with 5% non-fat dry milk in PBS with 0.05% Tween20 (PBS-T) and 5% goat serum. Blots were incubated with primary antibodies overnight at 4 °C. For verification, we used the following primary antibodies: mouse anti-human beta-actin (Sigma-Aldrich, Deisenhofen, Germany; dilution: 1:7000), rabbit anti-human RPE65 (Santa Cruz, Heidelberg, Germany; dilution: 1:800), anti-equine CRALBP antiserum (self-made; dilution: 1:1000) [7]), rat anti-human CD49c (Ascenion, München, Germany; reported cross-reactivity to horse, dilution: 1:500) and rabbit anti-human Fibronectin (Thermo Scientific, Schwerte, Germany; dilution: 1:1000). After washing, blots were incubated with respective peroxidase-coupled (POD) secondary antibodies (goat anti-rat IgG POD, Sigma-Aldrich, Deisenhofen, Germany; dilution: 1:3000; goat anti-rabbit IgG POD, Sigma-Aldrich, Deisenhofen, Germany; dilution: 1:3000; goat anti-mouse IgG POD, Sigma-Aldrich, Deisenhofen, Germany, dilution: 1:5000; and goat anti-horse IgG Fc-POD, Biozol, Eching, Germany; dilution: 1:10000). Signals were detected by enhanced chemoluminescence (ECL) on X-ray films (Christiansen, Planegg, Germany). Protein expression was quantified after imaging the signals on a transmission scanner using LabScan 5.0 software (GE Healthcare, Freiburg, Germany, 2003) and densitometric analysis with ImageQuantTL software (GE Healthcare, Freiburg, Germany, 2003). Signals were normalized to beta-actin content after staining of lanes with monoclonal mouse anti-beta-actin antibody (Sigma-Aldrich, Deisenhofen, Germany; dilution: 1:7000) prior to statistical analysis. As Kolmogorov-Smirnov test showed that data were not distributed normally (p ≤ 0.05). Mann-Whitney test was applied for calculation of statistical significance (p ≤ 0.05). Statistical analysis was performed using Paleontological Statistics software (PAST, available online: http://folk.uio.no/ohammer/past/index.html, accessed on 3 July 2012).
4.6. Immunocytochemistry
For immunocytochemistry, native equine RPE cells were set onto glass slides immediately after obtaining these cells from equine eye globes. Equine RPE cells of passage-4 were trypsinized in order to detach them from the cell culture flask and then set onto glass slides. Glass slides were then fixed in ice cold acetone for 10 min, followed by blocking with 1% BSA in PBS-T and 5% normal goat serum for 40 min at RT to prevent unspecific antibody binding. For fluorescence labeling, slides were incubated overnight at 4 °C with the following primary antibodies: mouse anti-human CRALBP (Santa Cruz, Heidelberg, Germany; dilution: 1:50), rabbit anti-human RPE65 (Santa Cruz, Heidelberg, Germany; dilution: 1:200), rabbit anti-human Fibronectin (Thermo Scientific, Schwerte, Germany; dilution: 1:500) and rat anti-human CD49c (Ascenion, München, Germany; reported cross-reactivity to horse; dilution: 1:200). Glass slides were then incubated with secondary antibodies for 30 min at RT (Alexa Fluor 546 goat anti-mouse, Alexa Fluor 546 goat anti-rabbit, Alexa Fluor 546 goat anti-rat; Invitrogen, Karlsruhe, Germany; dilution: 1:500). Cell nuclei were counter-stained with 4′,6 Diamidino-2-phenylindole (DAPI; Invitrogen, Karlsruhe, Germany; dilution: 1:1000). Finally, glass slides were mounted with glass cover slips using Dako fluorescent mounting medium (Dako, Hamburg, Germany). Fluorescent images were recorded using Axio Imager M1 or Z1 (Zeiss, Göttingen, Germany) and Axio Vision 4.6 software (Zeiss, Göttingen, Germany).